Latest / 1 More Paper / “Generative AI literacy across education and business: competencies, obstacles, and benefits—a systematic literature review” — Reicho et al. (2026)
Transcript
- You’re listening to ““Generative AI literacy across education and business: competencies, obstacles, and benefits—a systematic literature review”,” by M. Reicho and colleagues. Published in 2026. Abstract. This systematic literature review analyses AI literacy, focusing on the required competencies for, the obstacles arising from, and the benefits of, Generative AI (GenAI) in the fields of education and business. The analysis uses the PRISMA 2020 methodology with data from the SCOPUS and ERIC databases. A total of 538 articles were identified; of these, 206 were included after the full-text screening phase. Of those 206, only 33% (education) and 29% (business) were based on empirical research, highlighting the predominantly conceptual state of research. Using a combination of inductive coding and GenAI (ChatGPT-4o) validation, we identified AI literacy as a multidimensional concept comprising technical competencies (e.g. algorithmic literacy and prompt engineering), personal and interpersonal competencies (e.g. adaptability and collaboration), and ethical and critical thinking competencies (e.g. awareness of bias and ethical reflection). While educational literature emphasised pedagogical applications such as adaptive feedback and inclusive curriculum design, business research focused on process automation and data-driven decision-making. Top three identified obstacles included hallucinations, ethics and plagiarism, which manifested differently in contexts such as student assessment and personnel selection. Addressing these challenges will require targeted training modules, ethical governance structures, and institutional support in the form of faculty development programmes or workplace reskilling initiatives. Top three identified benefits of GenAI literacy training are described as critical thinking, personalized teaching and learning and personalized feedback across sectors. Introduction. Artificial Intelligence (AI) has become integral to various aspects of contemporary life, from self-driving cars to translation to virtual augmented-reality avatars. Generative AI (GenAI) - a specific type of AI that produces text, images, videos, and other data—is increasingly playing a significant role in professional environments. While professional teams are making increasing use of GenAI, this is often without any strategy or dedicated planning. The integration of GenAI into professional contexts necessitates consideration of the practices and needs specific to diverse environments, such as legal, tax, accounting and government issues. In education, GenAI can facilitate personalised learning experiences, while in business it can transform production processes, business models, and products. An understanding of how to incorporate AI into education is particularly crucial, as future social discourse and the capacity to innovate will be substantially shaped by educational systems. Applications like ChatGPT have demonstrated that effective utilisation of GenAI requires users who can engage fully with the technology. An increasing number of professional groups using GenAI tools need to be competent in prompt engineering - providing instructions to speech-based AI systems in order to identify rules, automate processes, or define output parameters. One study revealed that while 86% of employees identified a need for upskilling with regard to GenAI, only 14% reported receiving relevant training. Despite efforts to promote a conceptual understanding of GenAI, previous research has failed to examine how GenAI concepts are communicated across different professional groups. This article examines the specific needs of education and business professionals, as both domains present unique challenges with regard to critical thinking, accuracy of content, privacy concerns, and intellectual property management. We address the necessity of fostering AI literacy across these fields, considering educators, students, decision-makers, and industry professionals. In educational settings, GenAI can establish personalised learning ecosystems through adaptive content generation, scalable tutoring mechanisms, and differentiated instructional approaches. GenAI can enhance pedagogical practices through assessment automation, curriculum development, and opportunities for language practice, while improving accessibility through translation services, multimodal content conversion, and asynchronous learning. Nevertheless, significant challenges persist, including concerns about academic integrity, socio-economic disparities relating to access, professional development requirements, and implications for critical thinking development. As GenAI becomes more ubiquitous, the lack of wider understanding and proficiency in GenAI technologies and their societal implications is becoming a pressing concern. In education, GenAI has been characterised as a “double-edged sword,” with research suggesting that it should be introduced purposefully, ethically, and responsibly in order to maximise benefits and minimise potential threats. A consensus has emerged that educators must develop essential AI literacy competencies. Similarly, studies of business contexts demonstrate that GenAI has rapidly transformed working practices and key competencies for business professionals. GenAI currently serves both operational and competitive functions. Customer interactions are enhanced through conversation interfaces, personalised marketing algorithms, and recommendation systems. Operational efficiencies are delivered via automated content generation, code production, and synthetic data creation; and creative applications span design automation, content marketing, and accelerated product development. Competitive advantage is achieved through sophisticated market analysis, predictive modelling capabilities, and strategic planning tools. The business sector, like education, requires improved guidance on the responsible integration of GenAI to enable it to address data privacy, bias mitigation, job security concerns, and automation bias, and to promote ongoing adaptation, professional development, and strategies to ensure human oversight of decision-making processes. Although existing research has documented the growing applications, benefits and challenges of GenAI in education and business contexts, there is still a lack of comparative analysis across sectors that examines how AI literacy, professional competencies and responsible integration strategies can be coherently aligned in both areas. To address professional needs relating to GenAI concepts, we explored two research questions: 1. “What specific professional competencies are required to enable educators and. business professionals to approach GenAI in an appropriate, safe, ethical and reflective manner?” 2. “How do these professional competencies differ between education and the business. sector, and what implications do these differences have for professional development?” We investigated these questions through a systematic literature review. The following sections present the background to our specific areas of interest, beginning with AI literacy. Background. AI literacy refers to the specific knowledge and skills that enable individuals to critically understand, evaluate, and use AI-driven technology in ethical and safe ways. Ng et al. (2021) split AI literacy into four aspects (knowledge, use, evaluation, and ethical issues), adapting classic literacy frameworks. Laupichler et al. (2022) and Southworth et al. (2023) define AI literacy as the understanding required for individuals to engage meaningfully in broader AI discourse and make informed decisions regarding its application and impact. Research emphasises the importance of language capabilities when using AI-driven large language models (LLMs), suggesting that engagement with these systems could transform users into active co-creators through prompt engineering - the formulation of instructional inputs that trigger AI responses based on patterns learned during model training. Prompt engineering is a promising concept, connecting AI literacy to practical and professional contexts. Critical thinking skills—including information literacy, media literacy, and digital citizenship—are essential for responsible prompt engineering practices. By contrast, criticisms have been voiced, for example by Ballantine et al. (2024), who have warned of the potential threats arising from AI task takeover, loss of core competencies on the part of humans, lack of curricula, and ethical and methodological issues such as the deliberate manipulation of GenAI results with prompts. However, additional research is needed to assess the benefits and risks of prompt engineering and to address potential biases and ethical concerns. Zhang et al. (2023), for example, highlight ethical issues throughout the entire life cycle of AI systems, including data security, responsibility, accessibility, transparency and trust. The following sections focus specifically on the professional demands of AI literacy with regard to GenAI in education and business. GenAI in education GenAI has become a priority for educators worldwide. Its integration into educational settings offers potential for significant transformation, encompassing personalised learning methodologies, enhanced inclusivity, and skills development. The usage of GenAI in education continues to expand, with AI already enhancing personalised learning through adaptive systems that tailor content to individual student needs and progress. GenAI also facilitates communication through virtual assistants and chatbots that support students and educators, and contributes to assessment and evaluation through automated grading and adaptive testing that measure student performance more effectively. Zhai et al. (2024) offer a critical position on the use of GenAI in educational contexts. They note that students may accept AI-generated content uncritically, which can lead to impaired judgement and reduced cognitive engagement. They further emphasise the potential long-term consequences for essential skills such as critical thinking, decision-making and analytical reasoning, and raise concerns about algorithmic bias, misinformation and ethical challenges. Others, such as Olojede (2024) warn against overestimating technological solutions that ignore the complexity of social and educational policy challenges. In the context of ‘techno-solutionism’, she is critical of the fact that GenAI is often introduced into educational institutions without reflection on the ethical implications and without consideration of the social contexts in which these technologies are used. A significant challenge for educational institutions is the absence of specific, practical guidelines on the utilisation of AI tools and the re-evaluation of written work, particularly as AI use is frequently unverifiable. The effective integration of AI into education systems depends on the successful achievement of AI literacy, including teacher and learner preparation and fostering responsible use and critical attitudes toward the limitations of AI. Laupichler et al. (2022) emphasise that AI education should extend beyond programming skills to include ethical considerations, digital citizenship, and interdisciplinary applications. For non-technical students, AI literacy courses should prioritise conceptual understanding and practical implications rather than complex technical skills. GenAI in business and industry AI studies focusing on the business sector have expanded in recent years. Contemporary business environments are increasingly adopting GenAI across a range of sectors, primarily aiming to enhance customer experience and optimise product and service offers. Research into GenAI’s impact on professional practices reveals that the most common applications include information research and ideas generation, business communications, and summarising and revising texts. Many companies face challenges with AI implementation, particularly with regard to ethical implications such as data privacy, bias mitigation, and the maintenance of interpersonal trust. Practitioners stress the importance of ethical integrity in GenAI adoption and are exploring ways to use AI while safeguarding confidentiality and ensuring data privacy. Wach et al. (2023) take a pessimistic view of the deployment of generative AI in business and industry, emphasising its potential to exacerbate ethical, operational, and strategic risks. They highlight the displacement of human labour, which could undermine organisational resilience and stakeholder trust. They also argue that the rapid adoption of GenAI without adequate governance can intensify economic inequalities and introduce systemic vulnerabilities in corporate ecosystems. In the critical light of ‘techno-solutionism’, researchers remark that technological solutions developed in the West are not universally applicable, as they ignore local knowledge, cultural contexts and social structures. This mirrors colonialism, and the persistence of colonial patterns of power, knowledge and control. The rapid advancement of GenAI technologies means companies must pursue continuous adaptation and professional development; this brings the challenge of anticipating future skill requirements and complicates the delivery of training programs. Employees are concerned about job security and changing responsibilities, while AI decision-support systems risk “automation bias,” leading to uncritical reliance on those systems’ recommendations. Companies need to address these issues and put strategies in place to maintain human oversight and quality decision-making. Policy makers have also drawn attention to the need for AI literacy. The European AI Act, for example, states in Article 4 that “providers and deployers of AI systems shall take measures to ensure, to their best extent, a sufficient level of AI literacy of their staff”. This requirement has implications both for companies and for educational institutions from the moment they begin offering GenAI tools. Despite attempts to promote conceptual understanding of AI literacy, research on communicating AI concepts to different professional groups remains limited. Therefore, this paper addresses GenAI literacy by exploring the impact of AI on education and business, and proposing strategies to bridge the knowledge gap and enhance the ethical and practical use of AI within these sectors. In the section below, we introduce the methodology and methods used in our investigation. Both the education and business sectors recognise the transformative potential of GenAI in enhancing efficiency, personalising experiences and driving innovation. However, they share concerns about ethical challenges such as algorithmic bias, data privacy and the risks of uncritical reliance on AI systems. However, their respective priorities and areas of impact differ. The education sector focuses on building AI literacy among students and educators, emphasising ethical awareness and the integration of AI tools into existing pedagogical frameworks. In contrast, businesses prioritise managing operational risks, workforce displacement and the need for continuous adaptation to evolving AI technologies, while also addressing economic inequalities and data vulnerabilities. These shared and distinct challenges highlight the need for interdisciplinary collaboration and the development of clear, practical guidelines for the responsible implementation of GenAI. Methodology and methods This study explores the professional competencies needed for educators and business professionals to engage with GenAI responsibly, while examining how these competencies differ between the two sectors and their implications for professional development. Therefore, we conducted our systematic literature review using the PRISMA 2020 framework, since it was designed to enhance the transparency, completeness, and accuracy of systematic review reporting. The methodology comprises a 27-item checklist spanning seven key sections (title, abstract, introduction, methods, results, discussion, and supplementary information) and a flow diagram documenting the information processing sequence throughout the review process. The framework emphasises reproducibility and standardisation, to enable subsequent researchers to evaluate the review’s quality and potentially replicate its findings. Using PRISMA 2020 for systematic reviews examining emerging technologies like GenAI enables researchers to establish foundational knowledge while identifying gaps in research, methodological limitations, and opportunities for future inquiry, thereby contributing to the development of more robust theoretical frameworks and practical applications across professional contexts. The flow diagram for our identification, screening, and inclusion of literature is shown in Fig. 1. The systematic review was undertaken between July and October 2024 on the basis of the SCOPUS and ERIC databases. We were interested in AI literacy in the fields of education, business, and industry. The search query identified title, abstract and keywords: (generative AI AND (literacy OR competences OR competencies)) AND (learning OR education OR educators OR business OR industry OR school OR prompt). Inclusion and exclusion criteria Inclusion and exclusion criteria were defined with the aim of identifying the manuscripts to be included in the review. The inclusion criteria were defined as “generative AI, AI skills, AI competences, AI competencies, AI literacy” in combination with “learning, Fig. 1 PRISMA 2020 flow diagram. (adapted from Page et al., 2021) education, business, industry, school, prompt, teachers, educators, business workers, industry workers”. The exclusion criteria were “outside 2022–2024, no school students, no pupils, no children”, as we were interested in professional workers and employees (see Table 1). For example, if the article investigated pupils’ or students’ AI literacy or AI use, we excluded it. If the focus of research was teacher education, we included it. The full list of all articles with coding schema and inclusion and exclusion decisions is available here. Analysis After identifying and reading the relevant titles, all results were systematically compiled into a comprehensive data sheet, sorted by the extraction fields AI literacy competencies; benefits after AI literacy training; obstacles with GenAI; and the key topics in education and business, focusing on the similarities and differences between the two professional sectors. To assess consistency, a post-hoc inter-rater reliability check was conducted on a random sample of 30 studies, which were double-coded independently by two reviewers across three coding domains (90 coding decisions in total). The reviewers reached agreement in 72 cases, corresponding to an overall percentage agreement of 80%. Any discrepancies were resolved through discussion to ensure the coding framework was applied consistently. The analysis took an inductive approach. Each column of entries was collated in detail and duplicate content removed. We then developed categories using a traditional ‘pen-and-paper’ method. Terms with similar meanings were grouped together and assigned to inductive groups. This process was conducted manually by the research team to capture the nuances of meaning and terminology. In addition to human-led analysis, we used GenAI (ChatGPT, version 4o) to cross-check our inductive categories. To ensure full transparency, please the link to the prompts is here, along with the GenAI-based answers and suggestions. Although the AI-supported analysis produced similar thematic clusters, we chose to retain our own descriptors (e.g. AI suggested: “Ethical, Responsible & Governance Competence”, whereas our descriptor was “Ethical and Critical Competencies”). Despite the significant overlap, our final categories were based on researcher-led interpretation rather than automated labelling. Given the significant presence of pedagogical competencies in the data, we decided that this category would be addressed in depth in the pedagogical section, rather than being incorporated into the general framework. This triangulated approach — manual inductive analysis, group reliability discussions and AI-supported cross-validation — ensured both conceptual depth and methodological precision. To ensure ethical use of GenAI in our analysis, we implemented three key safeguards: all initial coding was conducted by human researchers without AI involvement, ChatGPT was used exclusively to validate, not generate, our existing categories, and all AI outputs were manually verified to ensure they did not introduce bias or overlook interpretive nuances present in our human-led analysis. Using GenAI in the analysis phase enabled large datasets to be processed and nuanced patterns to be identified, potentially revealing hidden findings or simply enabling previous findings to be cross-checked. However, challenges such as algorithmic bias, a lack of transparency, and the lack of interpretability of GenAI outputs meant that human oversight was essential. GenAI did not identify relationships that had been overlooked by the human research group, but it did confirm the results of the human-led study. Careful consideration also had to be given to ethical issues, such as data privacy and equity, to ensure that GenAI was integrated into the research in a responsible manner. Selection of papers The search query yielded a total of 538 articles, 249 of which were derived from the SCOPUS database and 286 from the ERIC database (see Figure 1). Three articles were added manually because of their special focus on business and human resources. Prior to screening, 11 items were excluded for reasons of form, including the presence of duplicates. In addition, 58 articles were excluded because the required degree of access was not available. The titles and abstracts of 469 articles were then screened, and 139 articles were excluded based on the established exclusion criteria. A total of 330 articles were included in a full-text review. A further 124 articles were excluded on the basis of the exclusion criteria. The main exclusion criterion was a focus on primary or secondary students, or teacher trainees. The final number of articles included in the analysis of the literature review was 206. The procedure is shown in Figure 1. During analysis, the following areas of extraction were given priority: the professional target groups (industry, education, business), the size of the sample, the GenAI technology used, the competencies acquired, the benefits after training, the field, the barriers, and the main findings of the articles. Results and main findings We have analysed and organised our findings around the two guiding research questions. First, the results address what the literature reports about competencies that enable educators and business professionals to engage with GenAI in appropriate, safe, ethical and reflective ways. To respond to our second research question, we present how these competencies differ between the education and business sectors, and what these differences imply for professional development, while also highlighting obstacles and benefits reported across both domains. The majority of the 206 articles were drawn from the field of education (86%), suggesting a robust interest within education research on the implications of GenAI. Conversely, existing research in business and industry remains limited (14%). This highlights the need for more research in this area to address knowledge gaps and promote a comprehensive understanding of the implications of AI for both educational and business professionals. Our analysis shows that in education, only about 33% of the articles included empirical work, with 67% being conceptual or theoretical. In business and industry, only about 29% of the articles related to empirical research, while 71% were conceptual or theoretical. The majority of literature in both fields is therefore not empirically-based. This underlines that AI literacy is still an emerging field, with basic definitions and conceptual understandings in the process of being formed. There is a lack of robust, data-driven insight into how AI literacy is taught and incorporated/ experienced in real-world settings, meaning that the field may still be in an exploratory phase, with limited real-world application and evaluation so far. Findings from the literature review This section presents the findings from the extraction fields of our systematic literature review, organised by three thematic categories, namely AI literacy, benefits after AI literacy training, obstacles with GenAI and key topics in education and business. We defined these extraction fields prior to the analysis on the basis of the research questions for this article. The focus was on identifying the necessary AI literacy competencies in education and business and on the differences between the two fields. Data for the respective extraction fields were then collected. The extraction fields were analysed and interpreted using inductive category formation. AI literacy AI literacy was the first of the four categories identified in the literature review. The AI literacy competencies described in the business sector and in the education sector strongly aligned in their emphasis on the types of competencies that were sought. From this, we identified that professionals across both sectors generally apply three distinct types of AI competencies (see Table 2): technical competencies, personal and interpersonal competencies and ethical and critical competencies. However, this apparent alignment between the education and business sector masks asymmetries in emphasis and purpose. While both sectors identify on a surface level similar competencies, there are differences in the underlying rationals. Pedagogical transformations through AI literacy and optimisation efforts in business performance represent diverging professional epistemologies, the details of which we will describe now. Technical competencies This category captures the advanced technical skills required to comprehend, develop, and operate AI systems, including programming, computational thinking, and algorithmic literacy as foundational elements for creating and implementing AI-driven solutions. Across the literature, two main perspectives emerge. Some studies conceptualise AI competence primarily as technical development, emphasising the ability to build and configure systems, while others adopt a broader, application-oriented view that focuses on working with data, understanding machine learning and deep learning, and interacting with AI systems in practice. This divergence reflects competing understandings of AI competence as either advanced technical production or applied, data-oriented use. Although data comprehension, statistical analysis, and mathematical knowledge are widely considered essential, the literature remains divided. At the same time, the growing emphasis on prompt engineering shifts attention from coding and model development toward communication and interaction with generative AI systems (De la Torre Madueño, 2023; Haugsbaken & Hagelia, 2024). A cross-sector comparison further highlights both similarities and structural differences. In educational research, technical competencies are often embedded within the TPACK framework, linking AI-related skills to broader pedagogical integration. In contrast, business literature emphasises technical skills such as programming and algorithmic literacy as means to leverage AI for organisational performance, with a stronger focus on outcomes such as efficiency and financial gain. While both sectors recognise the importance of technical expertise, only education provides a structured framework for integration. The absence of an equivalent model in the business context represents a significant gap, with existing descriptions remaining largely conceptual and lacking empirical validation. This difference may reflect a stronger emphasis on agility and context-specific adaptation in business contexts. We interpret these findings as indicating that technical competencies exist on a spectrum ranging from clearly defined expert domains to more fluid, context-dependent competencies. Furthermore, the absence of a consistent reference framework in the business context could suggest that requirements are evolving more dynamically and with less standardisation compared to education. Alternatively, it may mean that, when we talk about ‘business’, industry practices vary far more than in education, meaning that having a unifying framework could be misleading, as it would fail to recognise the many professional practices. This highlights that AI literacy in the technical domain remains an ongoing process of negotiation and definition. Personal and interpersonal competencies This category encompasses a broad set of competencies that support professional, cognitive, and interpersonal functioning in AI-related contexts, enabling individuals to integrate AI across domains. Frequently cited competencies include adaptability, flexibility, collaboration, teamwork, communication skills, creativity, innovation, problem solving, and idea generation. Across the literature, two central tensions emerge. First, competencies are framed either as individual capabilities—particularly adaptability and flexibility in response to rapidly evolving AI technologies, or as dependent on collective and organisational support, such as collaborative environments, mentoring, and institutional structures. Second, they are conceptualised either as social and communicative, emphasising collaboration and the ability to explain AI-related work to diverse stakeholders, or as drivers of innovation and performance, highlighting creativity, idea generation, and problem solving. These tensions point to a broader ambiguity in defining personal and interpersonal competencies as either socially oriented or performance-driven, and as individual or collective in nature. Additional competencies, including digital and media literacy, assessment and feedback literacy, self-regulated learning, socio-emotional skills, and leadership further extend this category. While digital, assessment, and feedback literacies are mainly associated with the practical use and evaluation of AI tools, self-regulated learning and socio-emotional skills are linked to personal development and learner support, and leadership and strategic vision relate to managing organisational change and conflict. A comparative view across the education and business sector reveals both overlap and contextual variation. Personal competencies such as the ability to collaborate, communicate, and adapt to changes in a professional environment are considered relevant in both education and business contexts. However, while leadership, negotiation, and strategic vision are often associated with organisational and business settings, similar challenges such as negotiating appropriate AI use are also highly relevant in education. Overall, both sectors recognise that effective AI integration requires a combination of individual competencies and organisational capacity, although the balance between these dimensions remains contested. Personal and interpersonal competencies in acquiring AI literacy are inherently multidimensional and context-dependent, making them difficult to categorise. The identified tensions suggest that these competencies act as a bridge between social interaction and performance-oriented outcomes, and between individual agency and organisational structures. Ethical and critical competencies Ethical and critical AI literacy competencies refer to the ability to analyse, evaluate, and engage with AI systems in a reflective and responsible manner. Across the literature, there is broad agreement on their importance, although terminology differs between education and business contexts. Critical thinking emerges as the most frequently emphasised competency, followed by the ethical and responsible use of AI. A central assumption is that users should not accept AI outputs uncritically, but instead actively question recommendations and evaluate generated content. At the same time, a key tension arises between this expectation of critical evaluation and the widespread risk of overreliance on AI systems. Frequently discussed competencies further include bias recognition, fairness, decision-making, limitations, information literacy, fact-checking, and source verification. These competencies highlight that ethical engagement is closely tied to understanding how AI systems function, as recognising bias and making fair decisions requires insight into data processing and system limitations. Similarly, information validation and source verification are essential for assessing the reliability of AI-generated content, while data interpretation supports awareness of evidential foundations and broader societal implications. Communication with AI is also framed as an ethical practice, emphasising respectful and reflective interaction that safeguards human dignity and autonomy. In contrast, competencies related to reflective writing, social impact, transparency, human oversight, and overdependence are mentioned less frequently. Across sectors, both education and business recognise the importance of ethical and critical competencies, despite differences in terminology. The shared emphasis on critical thinking, bias awareness, and responsible use suggests a common understanding that effective AI engagement requires active evaluation rather than passive reliance. At the same time, the comparatively limited attention given to broader societal, governance-related, and long-term considerations, such as transparency, human oversight, and social impact, shows a potential imbalance in how these competencies are conceptualised across contexts. Our analysis indicates that ethical and critical AI literacy is primarily framed as an individual cognitive responsibility centred on evaluating outputs, while less emphasis is placed on systemic and societal dimensions of AI use. The identified tension between critical engagement and overreliance suggests that current conceptualisations may underestimate the practical challenges users face in consistently applying these competencies. Overall, this points to an emerging but still incomplete understanding of ethical AI literacy, in which immediate evaluative skills are prioritised over broader, long-term considerations of responsibility and governance. It is important to note that while research on AI use in education and business seems to be highlighting similar competencies, the purpose for why this is important and who ought to benefit from such competencies is different. Technical, personal and ethical competencies should ensure that individuals can engage with AI technologies in ways that are thoughtful, responsible, and aligned with the values of humanity. There is a clear difference whether this is to benefit the education of future generations or the success of a business. It should also be noted that there is a synergistic relation between the three competency domains; and this means that deficiencies in any one area can compromise overall AI literacy. Benefits after AI literacy training This section offers an analysis of the findings of the review regarding the potential benefits of AI literacy training. Personal development: communication skills and global competence AI literacy has been shown to improve communication skills and support engagement in multicultural environments. In the business sector, communication is mainly linked to working in diverse contexts, where professionals need strong interactional skills to navigate intercultural communication. Beyond language proficiency and the ability to produce context-appropriate texts, GenAI can support understanding of different communicative situations. It is described as helping users become more aware of social cues, etiquette, and empathy, although this assumes that simulated interactions can contribute to real-world intercultural communication skills. Such training can also strengthen individuals’ ability to create and engage with AI-gener ated content, preparing them to participate in digital and global environments. The competencies with the strongest empirical support in this category include increased confidence, personalisation and personalised feedback, improved self-assessment, greater autonomy, creativity, and innovative expression. In educational research, these benefits are mainly linked to personalised feedback and self-directed learning. Other competencies, such as improved communication skills, language proficiency, confidence, reflection skills, and cultural and global competence, are mentioned less frequently. We interpret these findings as indicating that communication and global competence are viewed as both outcomes of and necessary skills for engaging with GenAI effectively. The stronger focus on individual benefits such as autonomy and self-assessment, especially in education, contrasts with the emphasis on intercultural communication in business contexts, indicating different priorities across sectors. This suggests that the communication-related benefits of AI literacy are not distributed uniformly across sectors. Training programmes may therefore need to explicitly address intercultural and global dimensions if these outcomes are to be realised beyond the context of individual education. Professional development: administrative efficiency and strategic management The literature presents AI literacy training as a key driver of professional development, offering benefits for both operational and strategic efficiency. It is assumed that GenAI will support economic growth, employability and career advancement across sectors. However, a tension emerges between its role in increasing organisational productivity and the need for professionals to actively interpret and evaluate AI outputs in order to make informed strategic decisions. There is a strong focus on operational efficiency, including automating administrative tasks such as assessments and lesson planning, and improving logistics to reduce time and effort. GenAI is also considered to support cost and resource management by enabling more efficient allocation and providing scalable solutions. In business, these competencies are associated with reduced workloads and enhanced performance, whereas in education, they facilitate teaching strategies, personalised learning and adaptive planning. Overall, AI literacy is associated with immediate efficiency gains and the development of analytical skills. Less frequently mentioned competencies include assessment automation, academic analytics, curriculum alignment, and support for research and programming. The findings suggest that AI literacy in professional development is mainly framed around efficiency, while its strategic dimension is less emphasised. The tension between automation and critical evaluation indicates that achieving higher efficiency also requires strong informed (human) decision-making skills. The tension between automation and critical evaluation suggests that improvements in efficiency depend largely on robust human decision-making skills. This is a challenge that presents itself in different ways across sectors. For example, education prioritises pedagogical judgement while business emphasises strategic and operational oversight. Ethical awareness and critical engagement Critical thinking is the most empirically supported benefit. Individuals reflecting during AI training acquire the ability to navigate the ethical implications of AI, to use AI responsibly, to critically evaluate AI content and technology, to recognise the importance of inclusivity, and identify AI biases. Johnson et al. (2024) underline that GenAI training can help learners to increase their critical literacy skills and use information ethically. This proficiency is critical for fostering a balanced and responsible approach to AI technologies, ensuring that they are used to support equitable and culturally sensitive practices. Walter (2024) discusses ethical principles for learners, like the fair use of GenAI, reflection on use, identification of potential limitations and the recognition of hallucinations. This may prepare learners for the ethical challenges and cultural nuances they may encounter in globally interconnected environments. The most frequently mentioned skills in this category included critical thinking, evaluation of AI outputs, understanding ethics and using AI responsibly. Less frequently mentioned skills in this category were awareness of the social and cultural impact of AI, an improved understanding of the limitations and capabilities of AI, and ethical communication and human-centred AI interaction. This suggests that current approaches prioritise individual-level judgment of AI outputs over deeper reflection on long-term social and cultural implications. This uneven development is evident in both sectors. However, business literature shows a particularly limited engagement with societal and long-term governance perspectives compared to the educational literature. The three main categories of personal development, professional development and ethical awareness and critical engagement are shown in Table 3. Obstacles with GenAI The following section lists and explains the obstacles created by GenAI in the context of education and industry. Each paragraph tries to explain how different obstacles arise in the two fields. Bias and discrimination GenAI systems frequently reflect biases that are embedded in their training data, which can lead to the dissemination of inaccuracies and unfair Table 3 Overview of benefits after AI literacy training as result of the analysed literature treatment that particularly affect underrepresented groups. The literature consistently assumes that GenAI systems reproduce the biases embedded in their training data. This leads to inaccurate and discriminatory outputs, which particularly affect underrepresented groups. A major concern emerges between the growing use of GenAI for content creation and decision-making, and the fact that it is not neutral. GenAI is not considered to be objective or impartial because the training data and system tuning both reflect human choices and values. This creates a clear contradiction between the expectation of fair and efficient automation and the risk that GenAI will reinforce existing inequalities. Across sectors, bias is shown to operate in different ways. In business research, biased data distributions, sampling and labelling can exacerbate discrimination based on gender and race, and can embed systemic inequities in automated decision-making processes. Furthermore, decision-makers may overtrust AI rankings, particularly in HR recruitment, thereby further reinforcing biased outcomes. In education, bias is primarily associated with content generation and student assessment, where personal characteristics such as gender, race, and socioeconomic background may influence outcomes or gradings. Underlying all of these studies is the assumption that bias is unavoidable unless professionals actively recognise and address it. However, the consequences of bias differ between sectors, affecting individual learning in education and organisational diversity and innovation in business. The tension between the perceived efficiency of automated systems and their potential to reproduce inequalities highlights a gap between technological expectations and real-world outcomes. At the same time, the sector-specific consequences indicate that bias operates differently depending on context, requiring tailored strategies for mitigation rather than universal solutions. Access, equity, and inclusion The literature agrees that the digital divide significantly restricts equitable access to GenAI, especially for marginalised communities, due to variations in infrastructure, resources, and technological proficiency. A key tension arises from the rapid development of advanced GenAI systems and restricted access to them. The high cost of the most capable models means they are mainly available to those who can afford them, while delays in the development of detection and monitoring tools further widen this gap. Conversely, there is a strong belief that inclusivity can be enhanced by deliberately designing GenAI tools to address the needs of diverse users and promote equal opportunities. However, studies show clear differences across sectors. In business, limited access is attributed to development costs and commercial incentives that restrict access, whereas equity concerns are linked to biases and data privacy risks. In education, access barriers are more closely associated with infrastructure gaps, limited teacher training, and unequal resources. Equity challenges are primarily discussed in relation to algorithmic bias, the underrepresentation of marginalised groups, and the absence of intersectional perspectives. Conversely, inclusion is constrained by limited stakeholder involvement and an absence of co-design with underrepresented communities. These aspects highlight that access and inclusion issues are more influenced by economic, institutional and design decisions than by technology itself. The tension between rapid technological advancement and unequal access highlights a risk that innovation may deepen existing inequalities if not accompanied by deliberate governance and design choices. These findings indicate a shift from viewing inclusion as a by-product of technological progress toward understanding it as an outcome that must be actively designed and managed. Academic integrity and plagiarism The literature highlights academic integrity and plagiarism as significant issues when using GenAI in educational and professional settings. There is a clear mismatch between the growing enthusiasm for GenAI’s ability to produce convincing academic texts and the resulting increased risk of plagiarism and unauthorised use of intellectual property. This also challenges existing assessment practices and creates pressure to reconsider how learning is evaluated. In education research, these issues were primarily mentioned in relation to academic integrity and student assignments. In the business sector, however, integrity and plagiarism were primarily associated with unauthorised use and the need for company guidelines. At the same time, the literature suggests that these risks can be mitigated by taking a responsible and preventive approach to GenAI use. This shows that integrity in research and learning should be guided by core principles such as honesty, transparency, fairness, and accountability. Rather than relying solely on detection or control measures, Rasul et al. (2023) highlight the importance of fostering integrity and raising awareness of the potential impact of GenAI on learning outcomes. These findings suggest that academic integrity challenges associated with GenAI are not solely issues of misuse, but reflect a broader misalignment between emerging technological capabilities and existing norms of authorship, assessment, and ownership. The contrast between education and business contexts indicates that integrity is framed either as a pedagogical concern or as a matter of organisational regulation. This suggests that sustaining academic integrity requires not only new policies, but also a rethinking of how learning, authorship, and originality are defined. Hallucinations, misinformation, disinformation and transparency The analysed articles highlight hallucinations as a key limitation of GenAI, referring to responses that appear credible but are factually incorrect, regardless of the specific tool used. Addressing misinformation and ensuring transparency are essential to reduce these risks and maintain trust in GenAI-generated content. In education research, the spread of mis- and disinformation through GenAI and social media platforms poses a significant threat, requiring the development of digital literacy and critical thinking skills among learners. Additionally, introducing AI in educational institutions often necessitates consideration of specific institutional needs. In business research, obstacles to AI adoption include accountability, and the reliability of communication and data as well as transparency when introducing AI to (existing) dynamic and rapidly changing environments. These findings indicate that hallucinations and misinformation are not only technical limitations but also central challenges for trust and accountability in AI use. The differences between education and business contexts suggest that risks are addressed through distinct priorities—developing critical literacy in education and ensuring reliable, transparent communication in business. At the same time, both perspectives highlight that trust in GenAI depends on users’ ability to recognise and manage uncertainty in AI outputs. Human oversight and overreliance The potential of GenAI to increase productivity is also accompanied by concerns. Although GenAI tools can support learning and working processes, their convincing but often inaccurate results may lead to overreliance, potentially undermining critical thinking skills.. While GenAI can support learners, excessive use may lead to overdependence and reduce the human interaction that is vital for developing key social and emotional skills. A dominant assumption across the literature is that humans must remain responsible for making final decisions, particularly in situations that could have significant consequences for individuals or communities. These two obstacles play a significant role in both the education and business sectors. Overreliance seems to be a more pressing issue in education, where learners may use GenAI without reflecting on its potential limitations. In contrast, human oversight is particularly important in business, for example in the selection of employees. These sector-specific differences and similarities highlight the need for a balanced approach in which human judgment remains central, highlighting that effective AI literacy requires not only technical use but also sustained critical awareness and responsibility. Privacy, security, and data governance The large-scale accumulation, processing and storage of sensitive data by GenAI systems escalates privacy and security concerns for all professional areas. The literature highlights a strong tension between the growing use of GenAI and the risk of unintended disclosure of personal and sensitive data. Professionals in education and business may share data without proper consent, creating serious privacy and confidentiality concerns, particularly in medical contexts. At the same time, GenAI’s reliance on large-scale user data raises concerns about surveillance capitalism and the monetisation of user activity and intellectual property. Similarly, introducing GenAI applications in both the education and business sectors requires a comprehensive hybrid governance framework that incorporates privacy-preserving techniques, model accountability and cybersecurity measures. Both sectors face similar major regulatory requirements regarding privacy, security and data governance. Wellbeing, human development, and creativity A major obstacle to integrating GenAI into educational and business ecosystems is maintaining human wellbeing, creativity and personal development. A central tension emerges between the productivity gains that GenAI promises and the risk that these gains come at the cost of the very capacities, creativity, metacognition, social engagement, that make professionals effective in the first place. Boguslawski et al. (2025), for example, argue that GenAI may hinder the development of human metacognitive skills needed for effective programming, while Tang (2024) emphasises the importance of preserving human creativity across all industries. This points to a contradiction in the literature: GenAI is simultaneously positioned as a tool for enhancing human potential and as a threat to the deeper cognitive and social processes that underpin it. In education, this tension is particularly acute, as over-reliance may undermine the development of independent thinking and collaborative learning. In business, the drive for efficiency risks reducing space for the creative experimentation and interpersonal engagement that support innovation. Addressing this obstacle therefore requires not only limiting overuse, but actively designing AI-integrated environments that foster, rather than erode, these essential human qualities. Table 4 summarises the key obstacles associated with GenAI in education and business. The most empirically supported obstacles are hallucinations, ethical concerns and plagiarism and academic integrity. An analysis of the competency alignment reveals both convergences and gaps between educational and business sectors. While both sectors emphasise ethical and critical thinking competencies, educational frameworks prioritise pedagogical applications and learner development, whereas business contexts focus on operational efficiency and return on investment. This misalignment suggests that graduates may possess strong understanding of concepts but lack the practical prompt engineering skills or data governance knowledge that is needed in corporate settings. Equally, business training often underemphasises the critical thinking and ethical reflection skills that are cultivated in educational contexts. We are now able to combine the obstacles described above and the competencies needed to avoid them: Bias and Discrimination (ethical/critical), Access, Equity and Inclusion (technical and ethical/critical), Academic Integrity and Plagiarism (ethical/ critical), Hallucinations, Misinformation, Disinformation and Transparency (ethical/ critical), Human Oversight and Overreliance (ethical/critical, personal/interpersonal), Privacy, Security and Data Governance (technical), Wellbeing, Human Development and Creativity (personal/interpersonal). Our analysis of obstacle-competency relationships Table 4 Overview of obstacles with GenAI as set out in the analysed literature Table 5 List of top 10 GenAI benefits, competencies and obstacles sorted in empirical frequency reveals a hierarchical pattern in addressing GenAI challenges. Ethical and critical competencies emerge as a way to directly address six of the seven obstacles identified (bias and discrimination, access and equity issues, academic integrity concerns, misinformation risks, overreliance tendencies, and aspects of privacy governance). Technical competencies and personal/interpersonal skills play complementary but equally vital roles—technical competencies are essential for understanding system limitations and implementing security measures, while personal/interpersonal skills enable effective collaboration and maintain human creativity in AI-augmented environments. This distribution suggests that while all three competency domains are necessary, ethical and critical thinking capabilities represent the first line of defence against the majority of GenAI-related risks. The following chart (Table 5) summarises and quantifies the specific skills, benefits and obstacles, listing them according to their empirical support or frequency of mention. The benefits section identifies critical thinking, personalisation in teaching and learning, and engagement as the most significant dimensions. Hallucinations, ethical risks and plagiarism are the obstacles that are mentioned most frequently. Consequently, critical thinking, ethics and digital skills are at the top of the list of required AI skills. The thematic clustering of these three areas is described in the relevant chapters of this paper (see the chapters on AI literacy, the benefits of AI literacy training awhich includes technicnd obstacles with GenAI). Key topics in education and business The sections below highlight the key AI literacy topics analysed in the literature review, divided into issues for education and issues for business. Key topics in education GenAI in education is consistently framed as a driver of pedagogical innovation that reshapes teaching, learning, and instructional design rather than functioning as a standalone tool. GenAI-supported hybrid environments enable potentials of personalisation, adaptive feedback, and learner engagement, while curriculum and instructional design increasingly integrate GenAI into existing subjects rather than treating it as a separate competence. At the same time, concerns grow regarding over-automation or the erosion of the teacher’s role, highlighting tensions between efficiency gains and the preservation of human-centred pedagogy and professional judgment. This shift is closely tied to evolving teacher competencies, which position GenAI-related skills as embedded within pedagogical practice rather than isolated technical expertise. Assessment and learning practices are similarly being reconfigured through GenAI, particularly via adaptive feedback systems and personalised learning pathways that support student autonomy and self-assessment. However, these developments intensify concerns around academic integrity, as GenAI-generated student work challenges traditional notions of authorship and assessment validity. In addition, system-level barriers—including unequal access, limited infrastructure, and insufficient teacher training—restrict effective implementation and contribute to uneven adoption across educational contexts. These issues are compounded by risks of algorithmic bias in automated feedback and grading, which may reinforce existing social inequalities related to gender or socioeconomic background. Our analysis shows that AI literacy must extend beyond operational skills to include critical reflection on ethics, governance, and societal implications, particularly in relation to justice, autonomy, and inclusivity. At the same time, overreliance on GenAI raises concerns about the erosion of metacognitive skills, creativity, and social engagement, which are central to meaningful learning. Consequently, effective integration requires holistic approaches that balance productivity gains with the intentional design of learning environments that sustain independent thinking and human agency. Overall, the findings suggest that the success of GenAI in education depends less on technological capability than on the ability of educational systems to align pedagogy, ethics, teacher competencies, and institutional frameworks in support of core educational values. Key topics in business The analysis of this article highlights several key topics centred on how GenAI is integrated into organisational practice, workforce development, innovation, ethics, and decision-making. GenAI is primarily used to support routine and knowledge-based tasks such as writing, summarising, and reporting, positioning it as a tool that enhances efficiency while still relying on human strategic and emotional judgement. This requires employees to develop prompt engineering skills and an awareness of AI limitations to ensure effective use. At the same time, there is a growing emphasis on AI literacy as part of workforce development, with a focus on structured training and broader skill sets that include cognitive, emotional, and ethical competencies. Across these areas, a central tension emerges between narrow technical training and more holistic conceptions of workforce development. GenAI is widely framed as a driver of innovation and economic competitiveness, with the potential to reshape work processes and improve organisational performance. However, effective adoption depends not only on technical competence but also on critical awareness of AI’s limitations. In response, the literature calls for structured ethical frameworks to ensure responsible AI use aligned with organisational and societal values. Further concerns encompass regulatory compliance (e.g. GDPR), organisational policy gaps, and risks such as bias, lack of transparency, and socio-political consequences in areas like recruitment. Finally, human-AI collaboration is consistently framed as requiring maintained human oversight, particularly in high-stakes decision-making. While AI supports efficiency, overreliance is seen as a risk, reinforcing the continued importance of human judgment, emotional intelligence, and interpersonal processes in business practice. The key implication is that successful AI integration is not determined by technology alone, but by how organisations balance profit orientation and efficiency gains with critical literacy, ethical responsibility, and sustained human decision-making capacity. Discussion. This section revisits our initial research questions with broader theoretical implications, practical training designs, and the dynamics of cross-sector knowledge transfer. Revisiting our research question Two research questions underpinned this systematic literature review. RQ1: “What specific professional competencies are required to enable educators and business professionals to approach GenAI in an appropriate, safe, ethical and reflective manner?” Based on our analysis of the professional competencies required by educators and business professionals, we determined that AI literacy is portrayed as the interplay of technical competencies, personal and interpersonal competencies and ethical and critical competencies. Each is necessary: technical skills enable application, interpersonal skills foster human-AI collaboration, and ethical awareness ensures critical reflection. Deficiencies in one domain undermine the others, leading to risks such as bias, ineffective governance, or poor implementation. A balanced baseline across all three, with role-specific details, is essential for responsible and effective engagement with AI. Our obstacle-competency mapping reveals a hierarchical pattern in addressing GenAI challenges that has significant implications for training design. Our synthesis suggests ethical and critical competencies may function as foundational responses against six of seven major obstacles identified in our review: bias and discrimination, access and equity issues, academic integrity concerns, misinformation risks, overreliance tendencies, and aspects of privacy governance. Technical competencies and interpersonal skills serve complementary but equally vital roles—technical skills enable understanding of system limitations and implementation of security measures, while interpersonal capacities preserve human creativity and enable effective collaboration in AI-augmented environments. This hierarchical structure suggests that AI literacy programs should prioritise ethical foundations, establishing critical thinking and bias awareness before advancing to technical applications such as prompt engineering or data governance. RQ2: “How do these professional competencies differ between education and the business sector, and what implications do these differences have for professional development?” In both education and business, AI literacy is increasingly recognised as having multiple dimensions, encompassing not only technical skills, but also ethical awareness, critical thinking and the ability to collaborate with others. GenAI is associated with efficiency gains in content creation, feedback processes, and information management across both sectors, while also requiring professionals to engage with AI in a strategic and reflective manner. In education, these competencies are linked to adaptive instruction, personalised support, and inclusive learning environments. In business, they are associated with innovation, decision support, and operational performance. This shared emphasis suggests a broad consensus that AI literacy encompasses more than just tool use and involves socio-technical judgement. At the same time, however, AI literacy manifests differently across sectors, reflecting their divergent institutional purposes. In education, for example, it is closely tied to pedagogical transformation and the development of learners’ cognitive, ethical, and socio-cultural capacities. In business, however, AI literacy is primarily operationalised in relation to innovation, productivity, and competitive performance. While ethical concerns are present in both domains, they are framed differently: education emphasises humanistic and developmental concerns, whereas business often prioritises risk management and compliance. These contrasting orientations indicate that similar competencies are shaped by different purposes. This becomes particularly visible in content creation and personalisation. Although both sectors use similar technical capabilities, their purposes differ substantially. In education, AI-generated content and personalisation are directed towards learning outcomes, inclusivity and developmental appropriateness. In business, these same capabilities are often oriented toward market engagement and behavioural influence, raising concerns around persuasion, consent and autonomy. The implication is that AI literacy competencies cannot be understood as fully transferable generic skills, because the same technical function may require different ethical judgement depending on institutional context. In summary, while both sectors agree on the multidimensional nature of AI literacy, they diverge in their underlying motivation and goals and ethical orientations. Educational research literature on AI literacy dominates the analysed corpus with 86% of articles and foreground humanistic aspects and educational transformations. The research on business practices and AI was underrepresented with only 14% and were often conceptual rather than empirical. In business studies AI literacy is focused on efficiency of operations or risk management. We find this to be an asymmetry that is not easy to address since training programmes cannot simply use one sector programme to transpose to the other. Theoretical implications: towards an integrated AI literacy framework Taken together, our analysis describes AI literacy as the interplay of technical competencies, personal and interpersonal competencies as well as ethical and critical competencies. Our findings support a socio-technical model of AI literacy, in which ethical and critical competencies act as overarching competencies that govern the application of technical and interpersonal competencies. This redefines AI literacy as a layered capability architecture shaped by institutional context, rather than a bundle of skills. The competency hierarchy differs from other AI literacy frameworks in important ways. Sengsri and Khunratchasana (2024) defined AI literacy in terms of technical, business, and human competencies. However, this perspective neglects the ethical and critical thinking skills identified as essential in this study. Referring to Bloom’s Taxonomy, Ng et al. (2021) describe the mastery of AI literacy in four steps: Know and understand AI (acquiring fundamental concepts and skills); Use and apply AI (applying AI concepts ethically); Evaluate and create AI (performing higher-order thinking activities and enabling individuals to critically evaluate AI technologies, communicate and collaborate effectively); and AI ethics (considering human-centred and social responsibility issues). While this perspective differs slightly, it can be usefully linked to the descriptions in this article. Similarly, Laupichler et al. (2022) describe AI literacy as the ability to understand, use, and critically evaluate AI systems, as well as interact with them reflectively. This includes an understanding of how they work, their possible applications, limitations and ethical and social implications. Our hierarchical model extends these frameworks by explicitly mapping which competency domains address which obstacles, providing an evidence-based rationale for training sequencing rather than treating all competencies as equally foundational. The three-domain model presented in this review—which includes technical, personal/interpersonal, and ethical/critical competencies—should be understood as more than a taxonomy of skill sets, but as a structure showing interdependencies in which ethical and critical competencies are integrated meta-competencies. This architecture extends existing frameworks including Ng et al’s (2021) four-step model and Laupichler et al’s (2022) scoping review. This model provides an empirically grounded rationale for sequencing competency development. Ethical competencies should form the basis (and not be an add-on) for acquiring technical skills. This frames AI literacy to be not simply an acquisition of skills but a value-driven practice. The model is context sensitive, since specific ethical imperatives, institutional contexts and training tools differ substantially between the sectors. Generally, these findings support the idea that technology cannot be separated from an organisation’s rules and culture. They also move the AI ethics debate forward by arguing that ethical thinking should be a core requirement, not just an optional extra. Practical implications: sector specific training design Importantly, both fields recognise the need for structured training and institutional support to ensure the effective, responsible, and equitable deployment of AI tools, particularly in contexts where there are digital divides or skills gaps. Structured training modules differ in the two sectors. In education, programmes include: TPACK-AI integration workshops that combine technological knowledge with pedagogical applications, ethics-first curricula that start with bias recognition before introducing AI tools, and collaborative learning communities where teachers co-develop AI-enhanced lesson plans. In business, modules include: role-specific prompt engineering bootcamps for marketing, HR, and operations teams, scenario-based training using company data to practice bias detection, and cross-functional AI governance simulations that address privacy and compliance challenges. Furthermore, when designing AI literacy training for educational purposes, it has been suggested that a three-step framework can prove particularly effective: first, awareness-building through the understanding of terminology, natural language processing, ethics, and limitations; second, the development of practical skills including prompt writing, tool usage, and content evaluation; and third, the application of knowledge through hands-on problem solving and prototype development. It is argued that vocational training should follow similar principles, but emphasise immediate application in the workplace through the development of practical skills, hands-on experience, and job-specific competencies. For business sector courses, a four-dimensional framework proposed by Cardon and colleagues addresses the unique demands of corporate environments: Application focuses on understanding AI tools and aligning them with specific tasks; Accountability encompasses responsibility, content reliability, equity, and fairness considerations; Agency addresses decision-making rights and the risks of overreliance; and Authenticity preserves genuine communication and the essential human element in business interactions. These examples show that in education, training often focuses on developing a foundational understanding of AI and its ethical implications, as well as practical content generation and evaluation skills. Business training, on the other hand, often prioritises the practical, strategic and ethical use of AI tools including specific job roles and demands. In general, we found only limited references to design in the business sector in our dataset, compared with a large number of suggestions for the education sector. Put together, the findings suggest three overarching design principles for AI literacy training across both sectors: programmes should prioritise ethics, be sequenced from foundational critical awareness towards applied technical skills, and be sensitive to the specific goals and risks of each sector. These principles suggest that effective AI literacy training is not generic skills provision but principled competence development aligned with institutional purpose. Cross-sector insights and knowledge transfer However, cross-sector learning is useful and is most clearly identified by looking at shared obstacles and the different consequences across different sectors. Such obstacles include hallucinations, ethical concerns, threats to academic integrity through plagiarism, the reinforcement of bias and discrimination and the overreliance of GenAI-generated outputs. Access and equity remain major barriers, with marginalised communities having limited technological resources. Overreliance on GenAI may also undermine critical thinking, creativity, and social engagement. Additionally, issues around data privacy and governance highlight the need for clear regulatory frameworks. Ensuring responsible use of GenAI is essential to protect human development, foster inclusion, and maintain trust. The critical evaluation of GenAI obstacles demonstrates different consequences for both sectors. In education, biases and inequities could result in a generation of learners who are unprepared for a diverse world; in business, they could reinforce systemic inequities and thus hinder organisational growth and innovation. The erosion of academic integrity could diminish the value of educational qualifications, while in business it could have legal repercussions and lead to a loss of trust. Overreliance on GenAI can suppress creativity and critical thinking in both sectors, ultimately affecting human development and societal progress. The analysis of GenAI obstacles also reveals several important points of learning. Firstly, there is an urgent need for frameworks that prioritise ethical and critical thinking skills in order to address biases and discrimination. Secondly, targeted strategies must be developed to ensure equitable access to GenAI tools, particularly for marginalised communities. Thirdly, a culture of academic integrity should be fostered in education, while business should establish clear guidelines to prevent the unauthorised use of AI-generated content. Furthermore, it is crucial to enhance digital literacy and critical thinking skills to combat misinformation. Finally, a human-centric approach needs to be taken to the integration of GenAI, emphasising human oversight and creativity and ensuring that technology complements rather than replaces essential human skills. Key benefits of GenAI training include critical thinking, personalisation, personalised feedback, enhanced engagement and ethical understanding. Other benefits that are mentioned less frequently relate to confidence when using the tool, improved writing or communication skills, and self-assessment. The benefits after AI literacy training in education and business can be grouped into three main areas. Firstly, AI literacy training enhances personal development by improving communication skills and fostering intercultural competence, enabling effective engagement in diverse settings. Secondly, it supports professional development by increasing administrative efficiency, promoting strategic thinking, and enhancing employability through the effective use of AI tools. Lastly, it cultivates ethical awareness, preparing individuals to navigate digital environments responsibly and ensuring the equitable integration of AI technologies. These differences offer the business sector some insights stemming from education. For example, education’s emphasis on ethical awareness and critical thinking can help to mitigate automation bias and strengthen decision-making in business contexts. Its focus on human–AI collaboration and inclusivity can inform the design of AI systems that foster customer trust and satisfaction. Furthermore, pedagogical frameworks such as TPACK could inform the development of structured, role-specific AI literacy programmes within organisations. Furthermore, education’s personalisation approach, which prioritises learner welfare and data privacy, can inform more ethical business personalisation practices. Finally, collaborative models of AI-enhanced lesson design in education can provide a blueprint for cross-functional AI governance and innovation in business. Conversely, the business sector can offer valuable insights to the education sector. For example, the business world’s emphasis on productivity and measurable outcomes could help education move from conceptual discussions to the practical application of AI. The rapid adoption of AI for competitive advantage in business can provide models for more efficient integration into the curriculum. The use of real-world scenarios in business training could enhance hands-on problem solving and prototype development in education. Businesses’ role-specific AI training may also inform the design of AI literacy programmes aligned with diverse career pathways. Finally, the focus of businesses on innovation and market engagement could encourage a more entrepreneurial approach to integrating AI in education, including forming cross-sector partnerships to co-develop AI solutions. These cross-sector comparisons suggest that knowledge transfer in AI literacy occurs less through the direct adoption of competencies and more through the transfer of underlying principles. Although technical skills such as prompting, evaluation and personalisation may seem transferable, their ethical implications and correct application are influenced by institutional purpose. This implies that AI literacy competencies are only partially portable: it is not ready-made frameworks that transfer, but rather design principles such as human oversight, ethical reflexivity, and context-sensitive governance. Thus, cross-sector comparisons not only identify differences between education and business, but also highlight that AI literacy is a situated, socio-technical practice rather than a universal competency model. Conclusion. In conclusion, the findings of our systematic literature review underscore the critical importance of AI literacy across both educational and business sectors. We set out to conduct this analysis to improve guidance on responsible GenAI integration. As we navigate an increasingly AI-driven landscape, the necessity for a comprehensive understanding of AI technologies is becoming increasingly important. Our analysis reveals that AI literacy comprises a multifaceted set of competencies, including technical, personal and interpersonal, ethical and critical skills. These competencies are essential for educators and business professionals, and enable them to engage with AI in an appropriate, safe, ethical and reflective manner. There are great similarities between the AI literacy competencies identified in both the business and education sectors. This review highlights that while there are significant similarities in the competencies required across the two sectors – such as the need for digital literacy and the ability to critically evaluate AI outputs – there are also differences in their application. In educational contexts, AI literacy is closely tied to pedagogical innovation and the development of learners’ cognitive and ethical capacities. Conversely, in business environments, the focus is primarily on operational efficiency, innovation, and economic growth. This divergence reflects the unique goals and practices inherent to each field, necessitating tailored approaches to professional development and training. GenAI poses obstacles for both education and business, including bias, misinformation, threats to integrity, inequity of access, and data privacy risks. These issues can reinforce social and organisational inequities, erode trust, and undermine critical thinking and creativity. Robust ethical frameworks, equitable access strategies, and strong digital literacy initiatives are needed to address them. A human-centric approach that prioritises oversight, inclusion, and creativity is essential to ensure GenAI supports rather than stymies human development. Ultimately, the fostering of AI literacy is not merely an educational imperative; it is a societal necessity. As we prepare individuals to thrive in a digital and globalised world, it is crucial to cultivate a workforce that is not only technically proficient but also has a critical awareness of the ethical implications of AI. In doing so, we can harness the transformative potential of AI technologies while safeguarding human values and promoting inclusive practices. The path forward requires collaboration among educators, business leaders, and policymakers to create environments that support continuous learning and adaptation in the face of rapid technological advancement. This review reveals a critical empirical deficit in AI literacy research: only 33% of educational studies and 29% of business studies incorporated empirical data, suggesting that the field is still largely conceptual. Therefore, we can identify three research priorities that emerge from this gap that should inform future research agendas. First, robust empirical validation is essential. Longitudinal studies must track AI literacy outcomes across contexts and move beyond self-reported perceptions to show validated competency assessments. A particular focus should be on business contexts given their severe underrepresentation (14% of the total body of work) and the lack of theoretical frameworks that are comparable to the TPACK model used in education. Secondly, the integration of sectors demands systematic investigation. Research should examine how frameworks can balance the education sector’s emphasis on ethical reflection and learner welfare with the business sector’s priorities of measurable outcomes and innovation. It could be helpful to look at cross-sector partnerships that can offer opportunities to study how human–AI collaboration shapes trust, satisfaction and ethical practices in different organisational contexts. Thirdly, implementation research must address systemic barriers such as digital divides, resistance to adoption, and overreliance on AI through human-centric design approaches. Our examination shows that studies that integrate experiential learning, real-world problem solving and prototyping can shed light on how AI literacy translates from training into sustained professional practice. The limitations of the present study are twofold. Firstly, we did not use grey literature, reports or policy papers. Secondly, it is acknowledged that research is moving very fast, especially with regard to GenAI. The rapid development of this field may mean that the results of this study become outdated quickly. Acknowledgements. The authors utilized DeepL for language improvement and proofreading, and employed ChatGPT 4o to cross-check the categories of the inductive analysis presented in this paper. Authors contributions Michael Reicho, Irfan Jahic and Dominik Freinhofer were involved in the analysis, Michael Reicho and Kathrin Otrel-Cass contributed to theory and writing the article while Martin Ebner and Benedikt Brünner contributed in strategic planning of the content and feedback loops. We express our sincere gratitude to Christina Lechner for her valuable support and insightful contributions during the analysis of the literature. Funding This research has been funded by the Zukunftsfond Steiermark as part of the project “Prompt Engineering in Education and Business” in Austria. Data availability. The raw data is available under: the linked source. Declarations Competing interests The author(s) declare no potential conflicts of interest with respect to the research, authorship and/or publication of this article. Received: 17 June 2025 / Accepted: 4 May 2026