Latest / WTF! Women Talk Finance / EP 22: The Power and Ownership of AI with Dr. Armina Karapetian

Transcript

EP 22: The Power and Ownership of AI with Dr. Armina Karapetian · WTF! Women Talk Finance

• 0:00
  SPEAKER_01: I'm Jackie.
• 0:01
  SPEAKER_01: I'm Candace.
• 0:03
  SPEAKER_01: And this is WTF.
• 0:05
  SPEAKER_03: Grab your coffee, wine, water bottle, emotional support snack, no judgment, and let's get into it.
• 0:12
  SPEAKER_05: Hi, and welcome back to another episode of WTF Women Talk Finance.
• 0:17
  SPEAKER_05: We've got a really exciting episode today and an awesome guest.
• 0:22
  SPEAKER_05: Most of us already are using AI, sometimes intentionally and sometimes without even realizing it.
• 0:30
  SPEAKER_05: It may write your first draft, uh, summarizes your meetings, remembers a strategy, it knows maybe your client history or history on a project.
• 0:42
  SPEAKER_05: And increasingly, it may know more about how our businesses operate than any single person on the team.
• 0:52
  SPEAKER_05: It sounds incredibly powerful, but it also raises one hell of a question.
• 0:58
  SPEAKER_05: If AI is learning from our ideas, our decisions, our relationships, and our intellectual property, who actually owns what it knows?
• 1:08
  SPEAKER_05: And for women who are already navigating gaps in access, capital, authority, and recognition, does AI help close those gaps?
• 1:18
  SPEAKER_05: Or are we building the next powerful system without asking who benefits, whose work it absorbs, and who remains in control?
• 1:29
  SPEAKER_01: Our guest today lives at the intersection of those questions.
• 1:34
  SPEAKER_01: Dr.
• 1:34
  SPEAKER_01: Armina Karpishan is the founder and CEO of Iron Home AI, where she's building private AI technology designed to let organizations use artificial intelligence without surrendering their sensitive information to the cloud.
• 1:48
  SPEAKER_01: She's also in academia.
• 1:50
  SPEAKER_01: She's a teaching associate professor of marketing and holds the Charles and Ruth Hopping Professorship in Entrepreneurship at the University of Nevada in Reno.
• 1:58
  SPEAKER_01: Her research explores risk, decision making, and the ways social influence shapes our choices.
• 2:04
  SPEAKER_01: She also brings more than 20 years of experience across finance, marketing, strategy, and leadership, and has become a powerful advocate for women in entrepreneurship.
• 2:14
  SPEAKER_01: We are so lucky to have you here.
• 2:15
  SPEAKER_01: Thanks, Armina.
• 2:17
  SPEAKER_00: Thank you so much, Candace and Jackie.
• 2:19
  SPEAKER_00: Um it's great to be here.
• 2:20
  SPEAKER_00: Thank you for inviting me, and I'm really looking forward to our conversation.
• 2:24
  SPEAKER_05: Thank you.
• 2:25
  SPEAKER_05: We're we're so excited to have you.
• 2:27
  SPEAKER_05: We've spoken to you before, and it was such a good, really different perspective of the conversations we're seeing a lot about AI.
• 2:38
  SPEAKER_05: And today we're really having a conversation about trust and ownership and power and agency, which is a really important piece to the AI puzzle that I don't know is getting enough attention.
• 2:53
  SPEAKER_00: You're absolutely right.
• 2:55
  SPEAKER_00: So we no, no, AI is everywhere.
• 2:57
  SPEAKER_00: You can't escape.
• 2:58
  SPEAKER_00: But I feel like you can't, you know, open you know, your news feed and there's gonna be 10 stories about how it's bad and good and ugly and all in the same, right?
• 3:06
  SPEAKER_00: So it's it's definitely a force of nature that we're all trying to learn how to work with it and how to leave it, how to deal with this new technology that's changing our world very quickly.
• 3:18
  SPEAKER_00: And so from the organizational point of view, and that's how I look so at you know, AI adoption strategies and um AI implementations and the enterprise.
• 3:28
  SPEAKER_00: So what we see right now, it's almost like a catch-22.
• 3:33
  SPEAKER_00: So we see that AI becomes really useful.
• 3:36
  SPEAKER_00: And what you just mentioned, that you know, it knows so much about.
• 3:39
  SPEAKER_00: So it becomes really useful when it knows all this information about you.
• 3:43
  SPEAKER_00: Because generic AI, let's face it, AI slop is also everywhere, and we're all you know, drowning in the AI slop.
• 3:50
  SPEAKER_00: So the generic AI that's not trained on your data, uh, that just uses generic information is not particularly useful.
• 3:59
  SPEAKER_00: But on the other hand, once you start telling AI everything about your business, about your clients, your strategy, you know, you know, what you're thinking, you don't know where this data is going, right?
• 4:14
  SPEAKER_00: Uh so it is kind of this two forces.
• 4:17
  SPEAKER_00: That's what I see when we talk to companies, when we talk to founders.
• 4:20
  SPEAKER_00: So, yeah, so to be useful, it needs to know.
• 4:23
  SPEAKER_00: It needs to know context.
• 4:24
  SPEAKER_00: But once you start telling it the context, then that raises your data privacy risk.
• 4:30
  SPEAKER_00: And another kind of wrinkle on that whole story, why um some of the old true and tried um cybersecurity um you know measures are not really working very well, is before, if document from your organization was leaked out, you could look for this document.
• 4:51
  SPEAKER_00: That document does not exist in the AI world.
• 4:54
  SPEAKER_00: Once you put this information out, we don't know how it's gonna be learned, how it's gonna be repackaged, how where it's going, how it's gonna be used to trade.
• 5:03
  SPEAKER_00: So it's you can't really pinpoint what portion, you know, it's not anymore kind of an entity that leaked out.
• 5:10
  SPEAKER_00: It lives in some different format, so it's not so easy to identify.
• 5:14
  SPEAKER_00: That's why it so becomes in it as this additional level of challenge when it comes to data privacy.
• 5:23
  SPEAKER_05: How how do we navigate?
• 5:25
  SPEAKER_05: I and this is probably way too broad of a question.
• 5:29
  SPEAKER_05: It's just so darn easy to use, like you've mentioned.
• 5:34
  SPEAKER_05: It's so useful.
• 5:36
  SPEAKER_05: We all we all got an executive assistant, you know.
• 5:40
  SPEAKER_05: Um, and that executive assistant is getting better and better seemingly every single day, and more capable every single day.
• 5:48
  SPEAKER_05: And then a lot of it is because of how much we're putting into it, like you said, and how much we're willing to share.
• 5:54
  SPEAKER_05: Um but kind of like that, I don't know that I would get an executive assistant day one and then data dump all of my personal deep things to there would have to be this like period of building trust and seeing the trust play out.
• 6:13
  SPEAKER_05: And we're seemingly skipping that trust building step in general.
• 6:18
  SPEAKER_05: Is that what you're seeing?
• 6:20
  SPEAKER_00: Yeah, so I think it's not uh a bug, it's a feature.
• 6:23
  SPEAKER_00: So it is the the the chatbot.
• 6:25
  SPEAKER_00: So thinking the models are the models, right?
• 6:28
  SPEAKER_00: So it's the ways and math and all those calculations.
• 6:30
  SPEAKER_00: But what we interface with is the chatbot in most cases, right?
• 6:34
  SPEAKER_00: And it is has been designed to sound and feel like our confident, right?
• 6:41
  SPEAKER_00: So um the way that it talks to you, and it's just you in the screen, so it creates this illusion of a very intimate relationship, right?
• 6:49
  SPEAKER_00: So this is my you know trusted advisor who I can tell all my, you know, what I'm everything that I'm thinking about, uh, my plans for you know, again, uh my go-to-market strategy, whatever that might be, right?
• 7:05
  SPEAKER_00: So uh all my plans.
• 7:07
  SPEAKER_00: So it is has been designed to be that way, to feel that way.
• 7:12
  SPEAKER_00: And now we see actually a little bit of a pushback, and I think that some of the things that are happening to uh, you know, to meta, especially in a uh recent um ruling, in that kind of gives companies a little bit of a pause.
• 7:27
  SPEAKER_00: So the new tool, for example, that some of the companies are now rolling out for teenagers, they actually specifically design to sound less human.
• 7:37
  SPEAKER_00: Specifically for, you know, for children, for adolescents, there's a big risk that that relationship with AI is going to supplement or replace some of the human relationships.
• 7:49
  SPEAKER_00: So they actually now the company is trying to force certain markets and for certain populations to make it less human.
• 7:56
  SPEAKER_00: But you know, for us adults, anything else, like we're all taking responsibility for our actions, and um, we just need to be cognizant.
• 8:04
  SPEAKER_00: So I think uh maybe let's talk about a little bit risk, right?
• 8:07
  SPEAKER_00: So the risk perception, risk tolerance, and it's a topic that's very near and dear to my heart.
• 8:12
  SPEAKER_00: That's what I study, um, and that's what I do my research on.
• 8:17
  SPEAKER_00: So uh, and speaking of risk tolerance and risk aversion, women have kind of a more negative, I would say, um, association with counseling women are usually considered to be less risk tolerant.
• 8:32
  SPEAKER_00: And we all know high risk, high reward.
• 8:35
  SPEAKER_00: So uh in some senses, markets sometimes I feel punishes women for um you know this perceived uh less risk tolerance.
• 8:46
  SPEAKER_00: But I think that's completely we can kind of change the the rules of the game here because I think that women are not um risk-averse, we're risk-aware.
• 8:56
  SPEAKER_00: So, in a sense, we have a much keener or better understanding of the downside.
• 9:04
  SPEAKER_00: So there have been numerous studies that show that women are better long-term investors compared to men.
• 9:09
  SPEAKER_00: Women have better investing outcomes because men uh are prone to the overconfidence bias, right?
• 9:17
  SPEAKER_00: So men tend to trade much more than women.
• 9:21
  SPEAKER_00: And we know that you know the adage that uh trading is hazardous to your wealth.
• 9:25
  SPEAKER_00: We know that frequent trading leads to um worse results, worse investment results long-term compared to infrequent various.
• 9:33
  SPEAKER_00: So actually, I think that we need to change the conversation.
• 9:37
  SPEAKER_00: We need to stop talking about women being um, you know, risk-averse, but being risk aware.
• 9:44
  SPEAKER_00: We understand the downside.
• 9:45
  SPEAKER_00: And I think that, especially when it comes to investing, kind of we kind of went into finance somehow.
• 9:50
  SPEAKER_00: I don't know how we managed to end up here.
• 9:53
  SPEAKER_00: Um, it's important to understand how much risk you you're willing to bear, right?
• 9:59
  SPEAKER_00: It all sounds great to think what it's gonna have a 20%, 50% upside, but what happens if you lose money, right?
• 10:07
  SPEAKER_00: How would that make you feel?
• 10:08
  SPEAKER_00: And I think that very often we have this conversation thinking only about the upside, but really women have the better understanding of the consequences of the downside.
• 10:20
  SPEAKER_01: What does that look like within the AI landscape right now?
• 10:24
  SPEAKER_01: Especially you were talking about youth and and children.
• 10:27
  SPEAKER_01: I don't know if you were saying the the youth are sounding less human or the AI is geared.
• 10:34
  SPEAKER_00: So the AI targeted to AI targeted to children and youth is now designed to sound less human.
• 10:41
  SPEAKER_01: So from this lens of being women, being risk averse, and being more aware of the dangers, how do you how do you see this playing out right now with AI?
• 10:53
  SPEAKER_00: I think that we need to be our own advocates, right?
• 10:57
  SPEAKER_00: So we have um, you know, when you look at the landscape in our society, we have very powerful players in different parts of that landscape.
• 11:07
  SPEAKER_00: And those players shape AI strategies, the um, you know, policy with whatever, you know, little policy that we um have around AI.
• 11:16
  SPEAKER_00: But I think that we need to to what I kind of talk always about when I talk about AI, you cannot surrender your agency to AI.
• 11:26
  SPEAKER_00: Right?
• 11:26
  SPEAKER_00: So we need to keep our agency and not kind of feel you know defeated that this is technology that's being pushed at us, but be active participants in that conversation.
• 11:38
  SPEAKER_00: Because we need to have our voices heard and we need to be part of the conversation.
• 11:44
  SPEAKER_00: And I mean, every person kind of who uses AI actively will need to be aware of all the risks and uh how it's being implemented, um, how it's being you know used, how the data is being collected, who's going to, as uh Candace mentioned earlier, who owns this data, right?
• 12:03
  SPEAKER_00: So it's all very nebulous at this point.
• 12:06
  SPEAKER_00: And I think we need to have these conversations and put a lot of clarity around it.
• 12:12
  SPEAKER_05: You bring up some really interesting points about agency.
• 12:16
  SPEAKER_05: And I I want to kind of tie it into even the reframing of women over history is always really interesting.
• 12:24
  SPEAKER_05: And you talk about like we we've been painted it as risk averse when really actually risk aware is more appropriate, and you point out great great evidence for why there's plenty of you know historical commentary about women not being strong leaders, yet we actually have very strong statistics pointing out women make very strong leaders of countries, of businesses, um, and so forth.
• 12:49
  SPEAKER_05: And so we we the this narrative is interesting in having the agency within that, our own agency, and protecting that, it kind of speaks into the women have historically had their work borrowed or stolen or minimized or rewritten.
• 13:12
  SPEAKER_05: Is AI creating a new version of that problem?
• 13:20
  SPEAKER_00: So it is a complex question.
• 13:22
  SPEAKER_00: And unfortunately, we do know, and have been again many studies that show that AI is biased, right?
• 13:28
  SPEAKER_00: So basically, AI is like somebody put a mirror in front of our face.
• 13:32
  SPEAKER_00: So AI does not really invent anything new in that sense, it just basically gobbles enormous amounts of data and then builds models based on that data.
• 13:43
  SPEAKER_00: So whatever we have out there, that's what it's going to learn, right?
• 13:47
  SPEAKER_00: So we're teaching it um about our world, right?
• 13:51
  SPEAKER_00: So it's like you teach a child, right?
• 13:53
  SPEAKER_00: So this is the world.
• 13:54
  SPEAKER_00: So we give it some data points, but the data point, so whatever is in that data that we've produced so far, that's what the eye is going to basically gobble into its model.
• 14:03
  SPEAKER_00: And there is there is a lot of bias in the model, so we know that already.
• 14:08
  SPEAKER_00: Uh so from on one hand, it does allow um uh us as women to leverage that technology because it is a kind of uh equal level playing field, right?
• 14:19
  SPEAKER_00: So anyone can use this technology and um get you know really good at it and how to use it.
• 14:26
  SPEAKER_00: And um, I'm sure that's what you know, and I don't need to teach it, you know, preach that to you.
• 14:31
  SPEAKER_00: That's what you're doing in your businesses.
• 14:33
  SPEAKER_00: Um but at the same time, we need to be cognizant that this bias exists.
• 14:38
  SPEAKER_00: Um and the only way to, I think, overcome this bias is to produce a lot of data that shows the ad, that shows you know how women are great leaders, because AI in that sense is kind of um, you know, it's going to use the data that you give it.
• 14:52
  SPEAKER_00: So if you give it the data that shows that, you know, and you point to this data, then it's going to learn that in the next model, in the next integration model.
• 15:00
  SPEAKER_00: Because again, it just, you know, we kind of again still talk about AI as it's having its own mind.
• 15:06
  SPEAKER_00: And that sometimes uh I don't know.
• 15:08
  SPEAKER_00: I find, I don't know if you found that.
• 15:10
  SPEAKER_00: Um, like every day in the news, you read AI solved this problem, AI did that.
• 15:16
  SPEAKER_00: And I think what happened to the humans?
• 15:17
  SPEAKER_00: AI did maybe 20% of that.
• 15:20
  SPEAKER_00: But first of all, there are people who set up this problem who recognize that there is a problem that needs to be solved, right?
• 15:26
  SPEAKER_00: So there's some objective.
• 15:28
  SPEAKER_00: Um they set up this problem in the way that it could be solved.
• 15:34
  SPEAKER_00: They prompted the AI, AI did its part of the work, but then there was somebody else, some humans, who had to validate and very verify that the solution was correct and communicate it to the world.
• 15:46
  SPEAKER_00: So, in that whole workflow, when you think AI is only taking maybe doing 20% of the work, which is great computational work that you know maybe would have taken us much longer.
• 15:55
  SPEAKER_00: But the stories that we read, it's all about AI.
• 15:58
  SPEAKER_00: And I feel like it's I find it fascinating that humans almost almost disappeared from this headline.
• 16:04
  SPEAKER_00: It's like AI did no, it didn't.
• 16:05
  SPEAKER_00: You know, there were humans all around it, and it was part of the process, but not the whole story.
• 16:11
  SPEAKER_01: I love that.
• 16:11
  SPEAKER_01: And that's definitely the reality right now.
• 16:14
  SPEAKER_01: I think about a future where it may be possible to weigh the data.
• 16:19
  SPEAKER_01: And so you still have humans in the loop helping the algorithms giving weight to certain data such that I wonder if we could eliminate the bias over time through weighing data.
• 16:32
  SPEAKER_01: Whereas as humans, we we can't ever really remove those biases.
• 16:36
  SPEAKER_01: So that I'm I'm taking the lens of like women in the past have been excluded on so many different bases.
• 16:45
  SPEAKER_01: I'm wondering if AI can become more of an equalizer for women and and for people.
• 16:51
  SPEAKER_01: I'm in law, I think about judges, right?
• 16:54
  SPEAKER_01: Judges are humans, they're inherently biased.
• 16:56
  SPEAKER_01: They're not supposed to be.
• 16:57
  SPEAKER_01: We idealize them as neutral.
• 16:59
  SPEAKER_01: But what if an AI judge that truly has um bias eliminated through the weighing of data could be a future where um where we could get more an equal playing field in in several different areas?
• 17:17
  SPEAKER_00: Uh that's a great example.
• 17:18
  SPEAKER_00: So we know, for example, also from research that uh judges are usually more lenient after lunch, right?
• 17:23
  SPEAKER_00: So again, we're all human, right?
• 17:25
  SPEAKER_00: So we do have our um yeah, shortcomings as humans, but um I think that we do possess a moral compass that currently AI does not have.
• 17:36
  SPEAKER_00: Um and I think again, it is important to um, you know, I do not envision a happy future where the society is governed by the models.
• 17:51
  SPEAKER_00: I think it needs to be some sort of uh so AI is only good as it works within uh content and the context that humans provided it and for specific purposes.
• 18:01
  SPEAKER_00: And again, going back to the uh kind of the enterprise and the industry example, so we see a lot of failure in trying to implement or introduce AI into the um you know everyday workflows.
• 18:14
  SPEAKER_00: And uh I think that one of the reasons for that is that, first of all, there's some pushback from employees who are concerned that this model is going to now come or this tool is going to come and replace them.
• 18:24
  SPEAKER_00: But I think also um, you know, we need to recognize that we need to be again, we need to retain our agency and um use AI as a maybe thought partner, collaborator, but never give out uh you know our power and our agency.
• 18:41
  SPEAKER_00: You know, when AGI comes, that's a completely different story, and then we'll have a different conversation about it.
• 18:46
  SPEAKER_00: But I think that the current level of development, um, I think you know, we is we have to stay at the helm and um you know lead this conversation, not just uh let the um run.
• 18:57
  SPEAKER_00: But going back, Jackie, to your point uh about uh removing the bias, again, um so the you know, with um um the the there are so there are ways how uh those models how this bias can be adjusted by basically changing the weights and um in the models.
• 19:15
  SPEAKER_00: Uh so there there are ways to do it um to make AI kind of uh you know less biased.
• 19:23
  SPEAKER_00: Uh but I again we need to be a part of that conversation because we need to talk about that, we need to uh bring it to the attention that that it needs to change, right?
• 19:34
  SPEAKER_00: So um yeah, AI is just a mirror.
• 19:38
  SPEAKER_05: Are you seeing a gender gap in AI adoption?
• 19:44
  SPEAKER_05: I'm seeing a gender gap in the AI communication as far as replacing certain genders jobs, but are you seeing a gender gap in the adoption of AI?
• 20:00
  SPEAKER_05: And it could go back to women being more risk aware um and and cautious about sharing or you know, giving up big pieces of information.
• 20:14
  SPEAKER_05: Um just curious for with your take on that.
• 20:19
  SPEAKER_00: So I think this question is more of a systemic question.
• 20:22
  SPEAKER_00: So AI adoption, um, again, um I think that when it comes to just adapting the, you know, using um a chatbot, I think that women are as big of a user as men.
• 20:34
  SPEAKER_00: I don't I don't see any gender gap there.
• 20:36
  SPEAKER_00: Actually, I see uh women sometimes are more um kind of curious about this new technology and willing to give it a try and see how that works and experiment with it.
• 20:47
  SPEAKER_00: Um I think uh the where the the where there is, I think, a gap is that when it comes to more uh deeper implementation, and that's um comes from kind of an old uh systemic bias that uh there are the women in technology, right?
• 21:03
  SPEAKER_00: So just if you understand technology, then you understand how it works.
• 21:06
  SPEAKER_00: So I think it's not a new bias, it's just old bias that's because it is a new technology, and if you're trying to understand what's behind it, what's behind the chatbot and how it works, so they need to have kind of more technical knowledge, and historically there were more men in STEM than women, which uh kind of one of my kind of things.
• 21:26
  SPEAKER_00: I'm trying to get more women in STEM, uh, one of my goals.
• 21:30
  SPEAKER_00: Um so but I think in terms of the using the technology, like using the built the tools that are available on the market, I don't see any gap there.
• 21:39
  SPEAKER_00: I think women you know using as much as men.
• 21:43
  SPEAKER_05: In your role as an educator, are you seeing the younger generations evening that gap in STEM out?
• 21:54
  SPEAKER_00: I think hopefully, I think we'll see.
• 21:57
  SPEAKER_00: It definitely there is um, I think more amps.
• 22:00
  SPEAKER_00: Emphasis on it.
• 22:01
  SPEAKER_00: I think there's more support.
• 22:03
  SPEAKER_00: It's not, you know, you know, 30, 40 years ago, it was very unusual to have uh, you know, women in STEM or getting uh an advanced degree or work in that area.
• 22:15
  SPEAKER_00: I think now it's becoming you know, it's an everyday occurrence.
• 22:20
  SPEAKER_00: Um so for the younger generation, I don't see the gender gap, but I was seeing it's actually quite interesting.
• 22:27
  SPEAKER_00: So there are um, you know, just looking at my like 20-year-olds who completely embrace the technology, but there is surprisingly there is a good proportion of my students who are afraid of it and don't even want to touch it.
• 22:42
  SPEAKER_00: So there is a resistance.
• 22:44
  SPEAKER_00: Uh, and I think that because that's the generation that's in the social media kind of play out and um all the um negative uh societal effects that it had on them, on their friends.
• 22:57
  SPEAKER_00: So they're very cautious of this technology.
• 22:59
  SPEAKER_00: Again, there's kind of two camps that, you know, some of them really into it and embrace it, but some really say, well, I don't really want to use it, and um, you know, kind of very cautious about it.
• 23:12
  SPEAKER_00: So I I found that actually interesting.
• 23:15
  SPEAKER_00: I I thought that they'll all be all over it, but that's actually not the case.
• 23:18
  SPEAKER_05: I find even younger kids, so my kids are much younger, um, their elementary school age, they're quick to call things out as AI.
• 23:28
  SPEAKER_05: I think we were worried they'd be more susceptible to believing what AI produces.
• 23:36
  SPEAKER_05: And I've noticed they're very critical of s they're like, oh, that's that's AI.
• 23:41
  SPEAKER_05: They can they can sense that something is not reality and they'll they'll say, Oh, that's that's AI, which is funny because I I do think there was this fear that that there would be this not knowing reality versus you know artificial, and the younger generation is quick to call it.
• 24:06
  SPEAKER_06: You can have a board, investors, advisors, and still have nobody you can actually talk to about what is really going on.
• 24:13
  SPEAKER_06: That is not unusual, that is the founder reality, and we step in right there.
• 24:18
  SPEAKER_06: I'm Tom Powell, and at the founder's office, we're proud to sponsor Women Talk Finance.
• 24:24
  SPEAKER_00: Yeah, so it's uh it's a great point.
• 24:26
  SPEAKER_00: So one thing that our brain, human brain, has evolved to do is to recognize patterns.
• 24:30
  SPEAKER_00: So I think initially when AI just came out, you know, it was just such a novel technology that um everyone just was mesmerized by the novelty, right?
• 24:38
  SPEAKER_00: It's just to have this amazing technology.
• 24:41
  SPEAKER_00: Um I think that again, and that goes back to AI, to my AI slop comment.
• 24:46
  SPEAKER_00: So AI slop comment, we can we can all recognize it, right?
• 24:48
  SPEAKER_00: We can recognize the patterns and how it's written.
• 24:51
  SPEAKER_00: Um and it's great that children that, you know, their brains, you know, working, you know, the it, you know, so they can recognize that.
• 25:00
  SPEAKER_00: My concern here is that this technology is going to evolve and that it's going to evolve very quickly.
• 25:05
  SPEAKER_00: Uh so two years down the road, they may get really good at um, you know, fooling even the most discerning eye and you know, um sounding really, really human and kind of becoming really advanced.
• 25:17
  SPEAKER_00: So it's definitely a catch-up game.
• 25:18
  SPEAKER_00: So at this point, I think that everyone's kind of saturated.
• 25:21
  SPEAKER_00: We've seen so much AI slop that we can recognize it when we see it.
• 25:25
  SPEAKER_00: But again, the technology is not uh is going to you know is going to improve dramatically in the next few years.
• 25:33
  SPEAKER_01: Armina, I want to switch gears a little bit.
• 25:36
  SPEAKER_01: Um, there's some overlap here.
• 25:39
  SPEAKER_01: You wear so many hats, and I'm curious how you came to be wearing these hats, meaning you are building a tech company featuring AI.
• 25:50
  SPEAKER_01: But like you said, the underlying is it's tech.
• 25:52
  SPEAKER_01: You're you're a woman in tech, and that's you know, science, technology, mathematics.
• 25:58
  SPEAKER_01: You are a huge advocate for women in the sciences.
• 26:01
  SPEAKER_01: And if I can give a little disclosure here, you run a company with your brother, which to me that sings to my heart.
• 26:07
  SPEAKER_01: My brother is my BFFFF.
• 26:09
  SPEAKER_01: So I I love that.
• 26:11
  SPEAKER_01: And then you've got this hat where you go and educate young people.
• 26:15
  SPEAKER_01: So I'm really curious, kind of loaded question, how you got here, and then kind of how you how you balance both of those or how you blend them, or rather, how you try to keep them separate, running a business and then educating.
• 26:29
  SPEAKER_01: I I'd just love for you to dive into that a little bit.
• 26:32
  SPEAKER_00: So there are definitely some interesting.
• 26:33
  SPEAKER_00: So my first good degree was in computer science.
• 26:36
  SPEAKER_00: So um I and that's how I started.
• 26:39
  SPEAKER_00: So I did program and see, did all those things in um you know early in my career, and then but very quickly moved more kind of the business side of things.
• 26:48
  SPEAKER_00: But it's I think that this um technology and stamps kind of colors how I think see things and kind of wants you to see the world a certain way from system thinking that uh remains for you for the rest of your life.
• 27:02
  SPEAKER_00: That's how we approach everything.
• 27:04
  SPEAKER_00: Um so I think that uh in both cases, so again, data privacy is something that um I've been you know thinking uh for a long time.
• 27:14
  SPEAKER_00: And I think that came from the realization kind of very early in my career that data is an asset.
• 27:21
  SPEAKER_00: Right?
• 27:21
  SPEAKER_00: So I kind of realized that very quickly that what companies working with, you know, Fortune 500 companies and uh different capacities, I quickly realized that what companies cared the most about was their data, right?
• 27:35
  SPEAKER_00: So the customer data, the internal data, so whatever that data is.
• 27:39
  SPEAKER_00: So I, you know, that's a result that data is an asset.
• 27:42
  SPEAKER_00: And then um when the um social media kind of came around, I was fascinated how willingly we were, you know, giving this data out with no payback, right?
• 27:55
  SPEAKER_00: You know, some just emotional or doping, you know, some um some different mechanism would trigger there in terms of the payback.
• 28:03
  SPEAKER_00: But companies were making a lot of money basically, if you don't pay for the product, you're the product, right?
• 28:08
  SPEAKER_00: So um that's you know, start, you know, started this whole kind of went to this path thinking about data privacy and data security, how it's important with AI, it became even kind of this these issues became, you know, they were amplified even more.
• 28:25
  SPEAKER_00: So because again, in this case, it just goes somewhere and we don't even know how it's been used, how it's been treated, how it's um what comes back, who owns this data, so it becomes this um, you know, the border is being definitely dissolved in front of our eyes.
• 28:42
  SPEAKER_00: Um so but I think that educational part, um, again, I never planned to be an academic.
• 28:48
  SPEAKER_00: That's not a part of the big plan.
• 28:51
  SPEAKER_00: Uh, but I come from a family where um, you know, of the educators and scientists, so it came very natural to me to be in the academia environment.
• 29:02
  SPEAKER_00: And teaching, I enjoy teaching more than now you know than I ever thought I would.
• 29:07
  SPEAKER_00: So I really so this part I you know just enjoy explaining, and we see somebody's, you know, students, the light bulb comes on.
• 29:15
  SPEAKER_00: That's you know, there's that's the dopamine high for me.
• 29:17
  SPEAKER_00: Um, but I was able to explain something in a way that makes sense to somebody else.
• 29:23
  SPEAKER_00: And now they've changed, like that, the whole worldview has changed.
• 29:27
  SPEAKER_00: So I think that's the you know, you know, teachers have more power than I think we um uh give them credit for.
• 29:33
  SPEAKER_00: So um so that part comes very natural.
• 29:36
  SPEAKER_00: But and starting the company, it was not it's not my first startup, so I was kind of always in the um entrepreneurship world in different capacities.
• 29:46
  SPEAKER_00: Um so starting the company again, my my brother, my co-founder, he thinks from a little bit different point of view.
• 29:54
  SPEAKER_00: So AI is a multiplier of the value of the company.
• 29:58
  SPEAKER_00: I come from the data security view, and that kind of there was a lot of synergy and um and now thinking.
• 30:04
  SPEAKER_00: So that's how the company, that's how that's why we founded the company.
• 30:08
  SPEAKER_00: Uh, and it's definitely, I cannot say that's like a part of my brain, it's just like teaching another part of my brain.
• 30:14
  SPEAKER_00: So but these are the things, it's like a Venn diagram.
• 30:18
  SPEAKER_00: So where things come together, I think about these things.
• 30:21
  SPEAKER_00: Um, I think about AI and you know the future of the technology or this particular technology and data.
• 30:27
  SPEAKER_00: So it's I still think about it, and then I apply it both in my teaching and um being the CEO of the company.
• 30:36
  SPEAKER_01: How are your students using AI, or how are you speaking with them about using AI?
• 30:42
  SPEAKER_00: So there are certain uh um, so the I'm so I think that we need to teach them how to use AI responsibly.
• 30:51
  SPEAKER_00: We cannot tell them, no, in my class, you can never use AI, because once they graduate and they go get their job, the employer is gonna expect them to use AI and know how to use it.
• 31:02
  SPEAKER_00: But what I tell my my students, if you just type a prompt in Chat GPT and give me the you know, whatever ChatGPT told you, why should I hire you?
• 31:14
  SPEAKER_00: I can do that myself, right?
• 31:18
  SPEAKER_00: Um, so you need to again apply your agency.
• 31:22
  SPEAKER_00: You um you have to be in the driving seat.
• 31:25
  SPEAKER_00: You cannot, you know, cede all your agency and your brain power to AI.
• 31:30
  SPEAKER_00: It's not the you know, it's I don't think it's a viable future for us.
• 31:34
  SPEAKER_00: Um so but it's a great tool to do brainstorming, to again be aware of the biases that it's always going to try to agree with you and say that everything that you're thinking is great.
• 31:44
  SPEAKER_00: So um be aware of that.
• 31:46
  SPEAKER_00: Always chase still will know that it hallucinates, it's you know, it gives the wrong information.
• 31:52
  SPEAKER_00: Um so you have to still check everything and think about it, but it's a great partner, and it can be a great partner.
• 32:00
  SPEAKER_05: Being your having your background in data security, but then also being a founder, being a serial founder, serial entrepreneur, in a startup, being an educator, having having a life, and passing this information on to future generations.
• 32:23
  SPEAKER_05: How are you personally using AI to support you in all of these ventures?
• 32:33
  SPEAKER_05: That's what I'm I'm wondering where my line is.
• 32:36
  SPEAKER_00: Okay, so um I've been I've been experimenting with some tools.
• 32:40
  SPEAKER_00: Again, the AI, the chatbot version of that that I use is uh the the private AI that our company has built.
• 32:47
  SPEAKER_00: Because um, you know, I if I need to find something that's I don't know what I'm gonna make it for dinner tonight, then you know, I have that in my refrigerator, and then you know, I haven't kind of probably also given more information to ChatGPT, but I don't really, you know, this generic information, I don't particularly uh worry about it.
• 33:05
  SPEAKER_00: Uh so I experimented with several tools.
• 33:08
  SPEAKER_00: Um uh I built some workflow automations and things like that.
• 33:14
  SPEAKER_00: Um and I found that um for certain things, if I need to quickly get some ideas on some something new that I'm trying that I'm thinking about, um just trying to brainstorm, that's where I use it the most.
• 33:35
  SPEAKER_00: Uh for the automation, I wasn't really impressed with what I was getting for the purposes that I was using it for.
• 33:41
  SPEAKER_00: So after using it for a few months, I was like, I can do better than this.
• 33:44
  SPEAKER_00: It's not the value wasn't there for me.
• 33:47
  SPEAKER_00: So um I think that when you know, the you know, there's agentic AI and there's the kind of more a chatboard, uh chatbot expert version of the AI.
• 33:58
  SPEAKER_00: For what I do, um agentic AI was not very useful.
• 34:03
  SPEAKER_00: But uh brainstorming partner um kind of are organizing my thoughts, um, so that's why I use it the most.
• 34:12
  SPEAKER_05: I personally love I love that you said that because I have been feeling behind because I have dabbled with a few automations and I haven't really loved the outcome, but I do use AI in the way that the tech bros seemingly tell me I'm behind the times, if I'm still using it in this way.
• 34:39
  SPEAKER_05: And I'm like, but I actually get a lot of value out of that.
• 34:42
  SPEAKER_05: I can I can riff with my chat on something that is seemingly low stakes as far as if it if it if it's utilized, if it's if it if it leaks, it's it's not high stakes information.
• 34:58
  SPEAKER_05: But I've gotten a lot of value out of that, and I can take that back to implementers in our team to expand on an idea.
• 35:08
  SPEAKER_05: Um and so I that was a really refreshing take from from you being being in the position you are, uh, because I have felt like I'm not building enough automations, I don't have enough enough bots working for me according to these bros on YouTube.
• 35:27
  SPEAKER_00: So that's you know another example of bias.
• 35:29
  SPEAKER_00: I think that you use it in a way that makes sense to you.
• 35:32
  SPEAKER_00: So I think that we've been told different stories, and I think that there are layers to the story, right?
• 35:37
  SPEAKER_00: So we've been told that this automation, again, you know, and it in some cases it does great things, right?
• 35:44
  SPEAKER_00: But for most people, I would say, even for most companies, we see again at the enterprise level, the AI adoption is very low.
• 35:52
  SPEAKER_00: So we're being told the story that how it's going to change the businesses, but the you know, the once you start looking at the data, it's not really there.
• 36:01
  SPEAKER_00: So, and I think that the reason because everyone starts, oh, we need to automate this and we need to automate this, and the employer is like, no, we're not doing that.
• 36:07
  SPEAKER_00: So we're not using it, it's not useful for me.
• 36:09
  SPEAKER_00: That's not how I do things.
• 36:10
  SPEAKER_00: Again, because AI doesn't know your business as well as you doesn't know and understand the context.
• 36:16
  SPEAKER_00: Uh, so I think that we shouldn't, as women, feel like, you know, you know, technobrosa saying that that's how you should be using it.
• 36:23
  SPEAKER_00: So you should be, you use it in a way that makes sense for your business, and that's where you can get value from the business.
• 36:28
  SPEAKER_00: Somebody else uses automation, great for them.
• 36:31
  SPEAKER_00: If they can figure out how to uh uh maybe automate some of the very routine processes, but in a creative business, and again, being a CEO or being a you know woman you know founder and doing more creative work, there's not a lot that I can automate.
• 36:50
  SPEAKER_00: I need to think myself, right?
• 36:52
  SPEAKER_00: So I use it as a research and a brainstorming tool.
• 36:55
  SPEAKER_00: So I don't do, and I never wanted to be in it.
• 36:58
  SPEAKER_00: Um actually, that was kind of my career.
• 37:01
  SPEAKER_00: One of my career goals was to kind of never to be in a position when I have to do the same things day in and day out.
• 37:07
  SPEAKER_00: So I prefer you know, I thrive in a more creative and loosely structured environment.
• 37:13
  SPEAKER_00: So for me, that's how I use it, and that's what makes sense for me.
• 37:17
  SPEAKER_00: I don't do a lot of repetitive tasks day in and day out that need to be automated.
• 37:22
  SPEAKER_01: I'm hearing that you use it to be creative, and I love that.
• 37:25
  SPEAKER_01: I also think I heard a rallying cry for a revolution in thinking for oneself, which I really like.
• 37:32
  SPEAKER_05: I just think that you described exactly my that can I I want I want the clip of that, and that's what I'm just gonna respond to whenever somebody's trying to pigeonhole me into like a structure.
• 37:45
  SPEAKER_05: My what I do most of the time is think creatively, and I have to have space and uh little loose structure to do that.
• 37:55
  SPEAKER_05: And I that is what I love using AI for is I do like having it as a it's an it's an available sparring partner that that doesn't seemingly have like a limit um of patience with me.
• 38:09
  SPEAKER_05: Um and it's not going to necessarily always get it totally spot on, but it might get me thinking in a new direction that I wasn't already thinking.
• 38:20
  SPEAKER_01: Let's make this really practical before we wrap up today.
• 38:24
  SPEAKER_01: What, Armina, in your opinion, are the three things every woman listening should stop putting into public AI tools immediately?
• 38:35
  SPEAKER_00: So any financial Yeah, it should stop.
• 38:38
  SPEAKER_00: So no financial data, uh no personal data, I would say.
• 38:43
  SPEAKER_01: Um Does your blood work count as your personal data?
• 38:47
  SPEAKER_00: It does.
• 38:47
  SPEAKER_00: So unfortunately, yes, unfortunately, and I know that there's a lot, you know, people using it for all sorts of purposes.
• 38:55
  SPEAKER_04: Um where's the line on personal data?
• 38:58
  SPEAKER_04: Yeah.
• 38:59
  SPEAKER_00: I need to know.
• 39:00
  SPEAKER_00: I have used it as my personal web and so actually the answer to that is very very simple.
• 39:06
  SPEAKER_00: Can somebody use this data to either try to uh target you or sell you something or discriminate you?
• 39:17
  SPEAKER_05: Okay, sell you something or discriminate.
• 39:20
  SPEAKER_05: Yeah, that would be my kind of I really like those that's easy.
• 39:26
  SPEAKER_05: That's Chris.
• 39:27
  SPEAKER_00: That's so easy to say.
• 39:28
  SPEAKER_00: I don't want to be a product like that, you know, as I say.
• 39:31
  SPEAKER_00: I I don't want to be become a just a product for somebody else that they're going to use it.
• 39:37
  SPEAKER_01: We've tried to buy another product.
• 39:39
  SPEAKER_01: I am the target market.
• 39:41
  SPEAKER_05: Yes, we've dislike that so there's this whole economy built on just shoving what the world thinks we need because now it knows us so well.
• 39:52
  SPEAKER_05: Um it's so it's infuriating actually.
• 39:58
  SPEAKER_05: In some ways it's convenient because you're like, wow, I this is going to make my life so much easier, but also gross that it knew that.
• 40:07
  SPEAKER_00: And it can be at my door by the afternoon.
• 40:11
  SPEAKER_00: Absolutely.
• 40:11
  SPEAKER_00: And especially with health data, and we don't have any um, you know, can insurance companies get a hold of this data, right?
• 40:18
  SPEAKER_00: So they say it is, and you know, can you, you know, can that be used against you?
• 40:23
  SPEAKER_00: Um, so that would be me concerned.
• 40:25
  SPEAKER_00: And you know they do say that they kind of anonymize the data, but there is big day, you know, again the thing that we don't know what they keep in and how they keep in it because again, it's very vague.
• 40:35
  SPEAKER_00: It's uh and again, it's not a bug, it's a feature because I think they're trying to figure out how they're going to monetize it because we all know the sort, you know, circle of the different topic, but circle of financing, we all know that some numbers are not really penciled out in terms of the ROI, uh, RI and an investment in AI.
• 40:52
  SPEAKER_00: So, how these companies to monetize it.
• 40:54
  SPEAKER_00: And uh my sense is that they're already kind of going down the path of the social media, which kind of used all the information to try it to sell you, right?
• 41:03
  SPEAKER_00: Um, but uh especially with health, like that that worries me that if any data can be used to um by insurance health insurance companies to discriminate and yeah.
• 41:15
  SPEAKER_05: That is that's very insightful.
• 41:18
  SPEAKER_05: I like those markers because they're very easy to follow.
• 41:21
  SPEAKER_05: So again, before you put it in to the chat, to Claude, to whatever AI you're using, can you be sold something?
• 41:30
  SPEAKER_05: Can it be used against you?
• 41:33
  SPEAKER_05: Can you be can you be discriminated against because of what you put in?
• 41:37
  SPEAKER_05: That's a really powerful uh point there.
• 41:42
  SPEAKER_05: So let's say somebody does want to brainstorm they they so women using AI uh to give them more of an even playing field in financial knowledge, in advice, making certain financial tools more accessible to women, is there a way for them to utilize AI safely to help them make financial decisions?
• 42:17
  SPEAKER_02: So you can use it for research.
• 42:20
  SPEAKER_02: Right.
• 42:21
  SPEAKER_00: So um if you want to research a particular company that you're thinking about investing, um you can use it for those purposes.
• 42:31
  SPEAKER_00: That's generic, that's publicly available in information.
• 42:34
  SPEAKER_00: So um I think that you know you can use that.
• 42:37
  SPEAKER_00: Don't tell that, you know, how much you're planning to invest or whether you're going to invest in it.
• 42:41
  SPEAKER_00: Just you know, you can ask, you make your question sound more generic.
• 42:45
  SPEAKER_00: Like, tell me about this company and help me understand their financials.
• 42:49
  SPEAKER_00: They just you know just publish the 10 Q 10K.
• 42:52
  SPEAKER_00: So can you can uh help me uh break it down and understand what it really means.
• 42:57
  SPEAKER_00: What's the what's in the um DNA section?
• 43:00
  SPEAKER_00: What you know, what are they, you know.
• 43:02
  SPEAKER_00: So you can use it for those purposes to do uh to do research, but again, I wouldn't rely on uh AI to uh help you make um investment decisions.
• 43:12
  SPEAKER_00: Again, um I come from a kind of very, you know, that's how I was you know taught at MIT.
• 43:19
  SPEAKER_00: So invest into uh an index fund and you stick with it, right?
• 43:23
  SPEAKER_00: So they know that there is uh systemic risk in the market and there's an idiosyncratic risk associated with a particular company.
• 43:30
  SPEAKER_00: Again, speaking about the risks, I want to minimize or be risk-aware.
• 43:34
  SPEAKER_00: So it could be it's a great stock that can go up, but what if it goes down?
• 43:38
  SPEAKER_00: So uh, you know, just talk to your financial advisor, a human being being who you trust.
• 43:44
  SPEAKER_00: Um, but you can use it to do research, definitely.
• 43:47
  SPEAKER_00: So if you want to do some research in the company, yeah, that's a great tool.
• 43:51
  SPEAKER_05: Does using a private AI tool like Iron Home AI does that is that a
• 44:00
  SPEAKER_00: viable option for those more sensitive maybe not loading in your your financial documents but or some of that more sensitive uh research so for for more sensitive research um you want to again I wouldn't trust any AI to give me a financial advice and uh again uh you can you know explain your risk and there are different metrics that we can you know that help you to um gauge you know your risk profile so I would start with that uh really uh when it comes to investing it's as much over it kind of sounds strange but it's as much as you learn about yourself as learning about the as learning about the market or a particular stock that you're interested in.
• 44:51
  SPEAKER_00: So understand know thyself would be my advice know what you can you know you know all the ups and downs look at the historical data again for that AI can be great.
• 45:01
  SPEAKER_01: So look at the historical data um you know what are the you know the returns on that particular stock or um index fund or you know whatever investing in for the last 10 20 what are the patterns um how is you know it is correlated with some other you know economic indicator so that's a great tool to do that but I think my kind of investment advice again would be start with yourself to start thinking you know what what is what are my goals what I'm comfortable with and uh no AI can tell you that only you know that Armina this I I know we were sitting under the umbrella of AI today and that's super fun but it's also really interesting that it becomes a conversation about trust and ownership and power and discretion and using your your own thinking capabilities and ultimately agency this word that you've used several times so thank you for having this conversation with us and for all of these all of these added elements in this AI conversation it's um it's exciting times I'd love for to to chat with you again in six months and see what's evolved and how much of this which I suspect is the majority holds true even as we as we navigate the the the nuances and the um evolution of the technique it's just it's been great talking to you Candace and Jackie I appreciate this opportunity you asked you know our conversation I thought that all the questions were really spot on and I really enjoyed it.
• 46:41
  SPEAKER_05: Thank you I hope our audience really enjoyed this and I I I think for us personally I loved getting from you these like kind of I I needed kind of some guardrails and you really helped clarify things that I had uncertainty about.
• 46:58
  SPEAKER_05: And that was a really and it's a very it was a very practical realistic approach.
• 47:03
  SPEAKER_05: So I'm so grateful to you.
• 47:05
  SPEAKER_05: So thank you for joining us today.
• 47:07
  SPEAKER_05: Thank you.
• 47:08
  SPEAKER_05: Thank you so much for listening and check in with us next week.
• 47:12
  SPEAKER_03: And like and subscribe oh yeah like and subscribe do those things okay that was today's episode of WTF if you laughed learned something or felt a little less alone make sure you hit follow and send this episode to a friend who might need it.
• 47:30
  SPEAKER_01: Women don't gatekeep especially not the good stuff.
• 47:33
  SPEAKER_03: We'll be back next week with more real talk more stories and probably more over sharing.
• 47:38
  SPEAKER_01: See you next time on WTF