Latest / USA Daily News / Charles Schwab CEO WARNS About AI Trading! Why Your Money Needs Protection

Transcript

Charles Schwab CEO WARNS About AI Trading! Why Your Money Needs Protection · USA Daily News

• 0:00
  Speaker 1: Artificial intelligence is beginning to reshape the way people interact
• 0:04
  Speaker 1: with financial markets, introducing a new generation of digital tools
• 0:08
  Speaker 1: that could change how investments are researched, managed, and traded.
• 0:13
  Speaker 1: For decades, investors have relied on brokerage platforms, financial advisors,
• 0:19
  Speaker 1: market analysts, and their own judgment to decide when to
• 0:22
  Speaker 1: buy or sell assets. Online trading made these activities faster
• 0:27
  Speaker 1: and more accessible. while mobile applications placed investment accounts directly
• 0:32
  Speaker 1: into the hands of millions of individuals. Now, the development
• 0:36
  Speaker 1: of autonomous AI agents is opening another chapter in this transformation,
• 0:41
  Speaker 1: raising the possibility that investors may eventually delegate more of
• 0:45
  Speaker 1: their financial decisions and transactions to software. This change presents
• 0:50
  Speaker 1: significant opportunities, but it also introduces difficult questions about risk, control,
• 0:56
  Speaker 1: and the responsibilities that come with allowing machines to act
• 1:00
  Speaker 1: on someone's behalf. To understand why this development matters, it
• 1:04
  Speaker 1: is important to distinguish between conventional AI assistants and AI
• 1:08
  Speaker 1: agents capable of taking action. Traditional AI-powered investment tools generally
• 1:14
  Speaker 1: help users gather information, summarize financial reports, explain market movements,
• 1:20
  Speaker 1: compare companies, or evaluate investment strategies. These tools can make
• 1:25
  Speaker 1: complicated information easier to understand, but the investor typically remains
• 1:30
  Speaker 1: responsible for deciding what to do with the information and
• 1:33
  Speaker 1: placing the trade. An autonomous agent can go further by
• 1:37
  Speaker 1: connecting multiple steps into a single process. Depending on its
• 1:41
  Speaker 1: permissions and technical capabilities, it could monitor financial information, identify
• 1:47
  Speaker 1: investments that match specified criteria, evaluate potential transactions, and submit
• 1:53
  Speaker 1: orders through an authorized trading platform. Instead of merely answering questions,
• 1:59
  Speaker 1: the system becomes capable of carrying out tasks that previously
• 2:03
  Speaker 1: required direct human involvement. That distinction is central to the
• 2:07
  Speaker 1: debate now emerging across the financial industry, because a tool
• 2:11
  Speaker 1: that provides information does not carry the same immediate financial
• 2:15
  Speaker 1: consequences as one that can move money and execute trades.
• 2:20
  Speaker 1: the attraction of these agents is relatively straightforward. Financial markets
• 2:24
  Speaker 1: generate enormous quantities of information every day, including company earnings,
• 2:30
  Speaker 1: economic indicators, interest rate decisions, corporate announcements, and changes in
• 2:35
  Speaker 1: investor sentiment. Monitoring all these developments can be difficult for
• 2:40
  Speaker 1: individual investors, particularly those who have jobs, families, and other
• 2:45
  Speaker 1: responsibilities outside the market. An AI agent could potentially observe
• 2:50
  Speaker 1: selected investments continuously, organize relevant information, and notify an investor
• 2:56
  Speaker 1: when specified conditions are met. With appropriate authorization, it might
• 3:01
  Speaker 1: also execute a predefined strategy without requiring the investor to
• 3:06
  Speaker 1: open an application and manually enter every order. Such capabilities
• 3:11
  Speaker 1: could reduce the time spent on routine investment tasks and
• 3:14
  Speaker 1: make certain forms of portfolio management more accessible to people
• 3:18
  Speaker 1: who cannot afford extensive professional financial services. Investors might also
• 3:23
  Speaker 1: benefit from systems that help them follow a consistent strategy
• 3:27
  Speaker 1: instead of reacting impulsively to every market headline. However, these
• 3:32
  Speaker 1: potential advantages depend on the quality of the technology, the
• 3:36
  Speaker 1: reliability of the information it receives, and the limits placed
• 3:40
  Speaker 1: on its authority. The appeal becomes even greater when considering
• 3:44
  Speaker 1: the difference between how markets operate and how individual people
• 3:48
  Speaker 1: make decisions. Stock prices can change rapidly in response to
• 3:53
  Speaker 1: unexpected announcements, economic data, or changes in expectations about corporate profits.
• 3:59
  Speaker 1: Human investors may struggle to process every relevant development, and
• 4:04
  Speaker 1: emotions such as fear, excitement, and overconfidence can influence their choices.
• 4:10
  Speaker 1: An automated system may be able to evaluate information more
• 4:14
  Speaker 1: consistently and respond quickly when its instructions are clear. It
• 4:18
  Speaker 1: could help an investor follow predetermined purchase limits, maintain a
• 4:22
  Speaker 1: target allocation, or review whether a portfolio has drifted away
• 4:27
  Speaker 1: from its intended balance. Yet, speed and consistency should not
• 4:31
  Speaker 1: be confused with sound judgment. AI systems can misunderstand context,
• 4:36
  Speaker 1: rely on incomplete information, or produce recommendations that appear reasonable
• 4:42
  Speaker 1: while overlooking a critical risk. A machine that makes a
• 4:45
  Speaker 1: decision quickly is not necessarily making a better decision, and
• 4:49
  Speaker 1: a system that follows its instructions precisely may still cause
• 4:53
  Speaker 1: harm if those instructions are poorly designed. Another important development
• 4:58
  Speaker 1: is the possibility that AI agents could alter the relationship
• 5:02
  Speaker 1: between investors and the platforms through which they access financial markets.
• 5:07
  Speaker 1: Brokerage firms have traditionally provided the systems that allow customers
• 5:11
  Speaker 1: to research securities, manage accounts, place orders, and monitor their holdings.
• 5:17
  Speaker 1: The growing role of artificial intelligence in financial markets has
• 5:21
  Speaker 1: prompted Charles Schwab Chief Executive Officer Rick Worcester to raise
• 5:25
  Speaker 1: an important question about the future of investing. How much
• 5:29
  Speaker 1: authority should individuals give to machines when their personal finances
• 5:33
  Speaker 1: are at stake? As AI agents become increasingly capable of
• 5:37
  Speaker 1: researching companies, evaluating market developments, and executing financial transactions, the
• 5:43
  Speaker 1: traditional relationship between investors and their brokerage platforms could undergo
• 5:48
  Speaker 1: a fundamental transformation. Worcester recognizes the potential of this technology,
• 5:54
  Speaker 1: but his position is clear. Investors should not give an
• 5:57
  Speaker 1: AI agent unrestricted permission to trade their money. Instead, they
• 6:02
  Speaker 1: should establish explicit guidelines that define how the system is
• 6:06
  Speaker 1: allowed to operate, what decisions it can make, and where
• 6:09
  Speaker 1: its authority must stop. His warning reflects a broader concern
• 6:14
  Speaker 1: within the financial industry that convenience must not come at
• 6:17
  Speaker 1: the expense of responsible decision-making and investor protection. Worcester's comments
• 6:23
  Speaker 1: come at a time when AI-powered financial services are attracting
• 6:27
  Speaker 1: growing attention from investors, technology companies, and established brokerage firms.
• 6:33
  Speaker 1: Platforms such as Robinhood and Coinbase are part of a
• 6:36
  Speaker 1: changing financial environment in which users increasingly expect technology to
• 6:41
  Speaker 1: perform tasks that previously required significant manual effort. AI agents
• 6:47
  Speaker 1: could eventually help individuals identify potential investments, compare opportunities, monitor
• 6:53
  Speaker 1: existing holdings, and carry out transactions according to predetermined instructions.
• 6:59
  Speaker 1: Rather than spending hours researching securities and entering orders themselves,
• 7:04
  Speaker 1: investors may be able to describe their objectives to an
• 7:07
  Speaker 1: automated assistant and allow it to handle parts of the process.
• 7:11
  Speaker 1: For customers seeking a simpler investment experience, the appeal is obvious.
• 7:16
  Speaker 1: Yet the ability to act independently introduces a level of
• 7:20
  Speaker 1: responsibility that goes beyond the functions of a conventional chatbot
• 7:24
  Speaker 1: or research tool. Once an AI system can execute transactions,
• 7:29
  Speaker 1: a misunderstanding or poorly defined instruction can have direct consequences
• 7:34
  Speaker 1: for someone's savings. Worcester's central argument is that investors need
• 7:38
  Speaker 1: to approach these systems with the same deliberate care they
• 7:41
  Speaker 1: would use when establishing an investment strategy through a traditional
• 7:45
  Speaker 1: brokerage account. A person using a trading platform today might
• 7:49
  Speaker 1: research a company, review its financial performance, determine an acceptable
• 7:54
  Speaker 1: purchase price, and submit a limit order specifying the maximum
• 7:59
  Speaker 1: amount they are willing to pay. These steps create a
• 8:02
  Speaker 1: process in which the investor retains control over the essential decisions.
• 8:07
  Speaker 1: An AI agent could potentially perform some of these tasks,
• 8:11
  Speaker 1: but it would still need clear instructions about the investor's intentions.
• 8:15
  Speaker 1: The goal should not be to eliminate human involvement entirely,
• 8:19
  Speaker 1: but to translate an investor's judgment into a defined process
• 8:22
  Speaker 1: that the system can follow. By establishing boundaries before granting
• 8:27
  Speaker 1: trading authority, individuals can make automation serve their financial objectives
• 8:32
  Speaker 1: instead of allowing a machine to interpret vague instructions in
• 8:36
  Speaker 1: ways they never intended. The distinction becomes especially important when
• 8:40
  Speaker 1: considering a seemingly simple request such as asking an AI
• 8:44
  Speaker 1: agent to buy stocks. That instruction leaves numerous questions unanswered.
• 8:49
  Speaker 1: Which companies should the system consider? How much money may
• 8:53
  Speaker 1: it invest? Should it prioritize established businesses or emerging companies
• 8:57
  Speaker 1: with greater uncertainty? What level of potential loss is acceptable?
• 9:02
  Speaker 1: Should the agent make a purchase immediately, wait for a
• 9:05
  Speaker 1: particular price, or seek approval before placing an order? Without
• 9:09
  Speaker 1: answers to these questions, an automated system would have to
• 9:13
  Speaker 1: interpret the request using its own assumptions, technical design, or
• 9:17
  Speaker 1: available data. Even if its decisions appear reasonable, they might
• 9:21
  Speaker 1: not reflect the investor's financial circumstances or tolerance for risk.
• 9:26
  Speaker 1: Worcester's warning highlights the danger of confusing broad permission with
• 9:31
  Speaker 1: meaningful control. Allowing technology to act is not the same
• 9:35
  Speaker 1: as ensuring that it acts in the investor's best interests.
• 9:39
  Speaker 1: A more responsible approach would involve building a structured set
• 9:42
  Speaker 1: of instructions before an AI agent begins executing trades. An
• 9:47
  Speaker 1: investor could specify a maximum amount for each transaction, limit
• 9:51
  Speaker 1: the proportion of a portfolio that can be placed in
• 9:54
  Speaker 1: a single company, and identify categories of investments that should
• 9:58
  Speaker 1: be excluded. The system could also be instructed to seek
• 10:02
  Speaker 1: approval before using leverage making unusually large purchases or changing
• 10:07
  Speaker 1: an established strategy More advanced arrangements might require the agent
• 10:11
  Speaker 1: to pause when market conditions become unusually volatile or when
• 10:16
  Speaker 1: information conflicts with the assumptions on which the investment plan
• 10:19
  Speaker 1: was based. These examples illustrate what rules of the road
• 10:23
  Speaker 1: could mean in practice. As artificial intelligence becomes more deeply
• 10:28
  Speaker 1: integrated into financial markets, One of the most important challenges
• 10:32
  Speaker 1: will be establishing clear rules that determine how much authority
• 10:36
  Speaker 1: an AI agent can exercise over an investor's money. The
• 10:40
  Speaker 1: technology may be capable of researching companies, evaluating market movements,
• 10:45
  Speaker 1: and executing transactions within seconds, but its effectiveness will depend
• 10:50
  Speaker 1: heavily on the instructions it receives. An investor who delegates
• 10:54
  Speaker 1: trading decisions without defining specific limits risks, allowing a system
• 10:58
  Speaker 1: to make choices that do not reflect their financial goals
• 11:02
  Speaker 1: or personal circumstances. This is why the idea of creating
• 11:06
  Speaker 1: rules of the road has become central to the discussion
• 11:09
  Speaker 1: surrounding autonomous investing. Before an AI agent is allowed to act,
• 11:14
  Speaker 1: there must be a clear understanding of what it can do,
• 11:17
  Speaker 1: what it must avoid, and when it should stop and
• 11:20
  Speaker 1: request human guidance. The first step is defining the investor's objectives.
• 11:26
  Speaker 1: Different people enter the financial markets for different reasons, and
• 11:29
  Speaker 1: those motivations should influence the way an automated system operates.
• 11:34
  Speaker 1: Someone saving for retirement over several decades may be willing
• 11:38
  Speaker 1: to tolerate short-term market fluctuations in pursuit of long-term growth.
• 11:43
  Speaker 1: Another investor who needs to preserve money for a house
• 11:46
  Speaker 1: deposit within a year may have much less room for
• 11:48
  Speaker 1: sudden losses. An AI agent should not treat these situations
• 11:53
  Speaker 1: as interchangeable simply because both individuals want their investments to grow.
• 11:57
  Speaker 1: Its instructions should reflect the investor's time horizon, financial priorities,
• 12:03
  Speaker 1: and willingness to accept uncertainty. Without this foundation, even an
• 12:08
  Speaker 1: agent that follows market data accurately could pursue opportunities that
• 12:12
  Speaker 1: are inappropriate for the person who owns the account. A
• 12:16
  Speaker 1: second essential safeguard involves establishing firm limits on how much
• 12:20
  Speaker 1: money the AI agent can commit to individual transactions. Investors
• 12:25
  Speaker 1: could specify the maximum amount available for a single purchase,
• 12:29
  Speaker 1: the total value of trades permitted within a day, or
• 12:32
  Speaker 1: the proportion of their portfolio that may be allocated to
• 12:35
  Speaker 1: one company or industry. These restrictions would help prevent an
• 12:39
  Speaker 1: automated system from concentrating too much capital in a narrow
• 12:43
  Speaker 1: range of investments. For example, if an agent interprets a
• 12:48
  Speaker 1: news report as a strong indication that a particular company's
• 12:51
  Speaker 1: shares may rise, it should not automatically be allowed to
• 12:55
  Speaker 1: place the investor's entire portfolio into that stock. A spending
• 12:59
  Speaker 1: limit creates a barrier between the agent's assessment of an
• 13:02
  Speaker 1: opportunity and the potential consequences of acting on that assessment.
• 13:07
  Speaker 1: It also helps ensure that the system cannot expand its
• 13:10
  Speaker 1: authority simply because it identifies what appears to be an
• 13:14
  Speaker 1: attractive trade. Rules should also define which types of investments
• 13:18
  Speaker 1: an agent is permitted to purchase. The financial markets contain
• 13:22
  Speaker 1: a wide variety of assets, from shares in established companies
• 13:26
  Speaker 1: to complex financial instruments that can expose investors to much
• 13:31
  Speaker 1: greater risks. Some products involve leverage, meaning that investors may
• 13:35
  Speaker 1: borrow money or use structures that magnify both gains and losses.
• 13:40
  Speaker 1: Others can be difficult to understand, expensive to trade, or
• 13:44
  Speaker 1: highly sensitive to market conditions. An investor who intends to
• 13:49
  Speaker 1: build a straightforward portfolio of diversified assets may not want
• 13:53
  Speaker 1: an AI agent to venture into highly speculative investments without
• 13:57
  Speaker 1: explicit permission. A properly designed system should therefore operate within
• 14:02
  Speaker 1: an approved investment universe and respect any exclusions the account
• 14:06
  Speaker 1: holder has established. Restrictions on particular securities, asset classes, or
• 14:12
  Speaker 1: trading strategies can help ensure that automation remains aligned with
• 14:16
  Speaker 1: the investor's preferences rather than pursuing every opportunity the system identifies.
• 14:22
  Speaker 1: Another important rule concerns the level of independence the agent
• 14:26
  Speaker 1: is allowed to exercise. Not every investment decision needs to
• 14:30
  Speaker 1: be treated in the same way. An investor might permit
• 14:34
  Speaker 1: an AI agent to monitor market conditions, gather information, and
• 14:38
  Speaker 1: recommend purchases without giving it permission to execute every transaction automatically. Alternatively,
• 14:45
  Speaker 1: the investor might authorize routine trades within a predetermined budget
• 14:50
  Speaker 1: while requiring human approval for unusually large purchases, substantial portfolio changes,
• 14:56
  Speaker 1: or decisions involving additional borrowing. This creates a graduated system
• 15:01
  Speaker 1: of authority in which the agent can perform relatively straightforward
• 15:05
  Speaker 1: tasks independently but must seek permission when the potential consequences
• 15:10
  Speaker 1: become more serious. Such an arrangement could preserve much of
• 15:14
  Speaker 1: the convenience of automation while keeping important financial decisions under
• 15:18
  Speaker 1: human supervision. It would also give investors a practical way
• 15:22
  Speaker 1: to expand or reduce the agent's responsibilities as they gain
• 15:26
  Speaker 1: confidence in its performance. Clear rules are equally important when
• 15:31
  Speaker 1: market conditions change unexpectedly. Giving artificial intelligence greater control over
• 15:37
  Speaker 1: investment decisions could make financial markets more accessible and efficient,
• 15:42
  Speaker 1: but it also introduces risks that investors may not fully
• 15:45
  Speaker 1: appreciate until something goes wrong. The ability of an AI
• 15:49
  Speaker 1: agent to analyze information, recognize patterns, and execute trades at
• 15:54
  Speaker 1: high speed can be valuable when the system operates within
• 15:58
  Speaker 1: clear boundaries. However, these same capabilities can magnify the consequences
• 16:03
  Speaker 1: of a mistake when the technology misunderstands an instruction, relies
• 16:08
  Speaker 1: on inaccurate information, or responds to changing market conditions in
• 16:12
  Speaker 1: an unsuitable way. Unlike a traditional investment assistant that simply
• 16:17
  Speaker 1: provides suggestions, an autonomous trading agent can potentially turn a
• 16:21
  Speaker 1: flawed conclusion into an actual financial transaction. When money moves automatically,
• 16:27
  Speaker 1: the time available for a person to identify and correct
• 16:30
  Speaker 1: an error may become much shorter, making appropriate safeguards essential
• 16:35
  Speaker 1: to protecting investors from avoidable losses. One of the most
• 16:39
  Speaker 1: immediate risks involves the possibility that an AI agent may
• 16:43
  Speaker 1: misunderstand what an investor wants to achieve. People often describe
• 16:47
  Speaker 1: financial goals in broad terms, asking for investments that offer
• 16:51
  Speaker 1: strong growth, opportunities to earn additional income, or ways to
• 16:56
  Speaker 1: improve the performance of an existing portfolio. These requests may
• 17:00
  Speaker 1: seem straightforward to a human financial advisor who can ask
• 17:04
  Speaker 1: follow-up questions and consider the client's wider circumstances. An automated system, however,
• 17:11
  Speaker 1: may interpret the same instructions according to patterns in its training,
• 17:15
  Speaker 1: the information available to it, or the specific rules governing
• 17:19
  Speaker 1: its operation. An instruction to maximize returns, for example, could
• 17:24
  Speaker 1: lead to decisions involving considerably greater risk than the investor
• 17:28
  Speaker 1: intended if acceptable losses and restrictions have not been clearly specified.
• 17:34
  Speaker 1: Even when the system is functioning as designed, the outcome
• 17:37
  Speaker 1: can be damaging if its interpretation differs from the investor's
• 17:41
  Speaker 1: actual expectations. This makes precise instructions and carefully defined limits
• 17:46
  Speaker 1: crucial whenever an AI agent receives permission to trade. The
• 17:50
  Speaker 1: problem becomes more complicated when the technology encounters incomplete, misleading,
• 17:56
  Speaker 1: or outdated information. Financial markets respond to a constant flow
• 18:01
  Speaker 1: of company announcements, economic reports, forecasts, political developments, and commentary
• 18:07
  Speaker 1: from investors and analysts. Some information is reliable, while other
• 18:12
  Speaker 1: material may contain speculation, errors, or deliberate attempts to influence
• 18:17
  Speaker 1: market perceptions. An AI agent that misinterprets a headline or
• 18:22
  Speaker 1: gives excessive weight to an unreliable report could identify an
• 18:26
  Speaker 1: investment opportunity that does not genuinely exist. It might then
• 18:31
  Speaker 1: execute a transaction before the investor has an opportunity to
• 18:34
  Speaker 1: assess the underlying facts. This risk is particularly important during
• 18:39
  Speaker 1: periods of uncertainty when rumors and rapidly changing expectations can
• 18:45
  Speaker 1: produce significant movements in asset prices. Even an advanced system
• 18:49
  Speaker 1: may struggle to distinguish a meaningful development from temporary noise,
• 18:54
  Speaker 1: especially when the information available is contradictory. Investors, therefore, need
• 18:59
  Speaker 1: to recognize that sophisticated analysis does not eliminate uncertainty and
• 19:04
  Speaker 1: that automated decisions should be supported by reliable data and
• 19:08
  Speaker 1: appropriate verification procedures. Software failures create another potential source of
• 19:13
  Speaker 1: financial harm. AI-powered trading systems rely on interconnected technologies, including
• 19:20
  Speaker 1: brokerage interfaces, data feeds, account permissions, and transaction processing systems.
• 19:27
  Speaker 1: A problem in any part of this chain could affect
• 19:29
  Speaker 1: the way an agent behaves. An incorrect data feed might
• 19:33
  Speaker 1: show a misleading price, a software defect could cause an
• 19:36
  Speaker 1: order to be submitted more than once, or a technical
• 19:40
  Speaker 1: interruption might prevent a system from responding properly to changing conditions.
• 19:44
  Speaker 1: Poorly designed instructions could also create a feedback loop in
• 19:48
  Speaker 1: which the agent repeatedly responds to the consequences of its
• 19:52
  Speaker 1: own earlier actions. These possibilities do not mean that automated
• 19:56
  Speaker 1: investing is inherently unsafe, but they demonstrate why financial systems
• 20:00
  Speaker 1: require testing, monitoring, and reliable ways to stop operations when
• 20:05
  Speaker 1: irregular behavior is detected. If a trading agent is permitted
• 20:10
  Speaker 1: to act continuously without meaningful supervision, even a relatively small
• 20:15
  Speaker 1: technical problem could develop into a larger financial incident. Speed
• 20:20
  Speaker 1: is one of the strongest selling points of autonomous AI agents,
• 20:24
  Speaker 1: yet it can also make errors more difficult to contain.
• 20:27
  Speaker 1: A person managing an account manually may pause between transactions,
• 20:31
  Speaker 1: review an order, or reconsider a decision after noticing new information.
• 20:36
  Speaker 1: An automated system can potentially submit multiple orders within a
• 20:40
  Speaker 1: short period if its instructions encourage repeated trading or if
• 20:44
  Speaker 1: it incorrectly interprets changing market signals. The emergence of autonomous
• 20:49
  Speaker 1: artificial intelligence agents is creating a new competitive challenge for
• 20:54
  Speaker 1: the financial services industry, forcing established brokerage firms to reconsider
• 20:59
  Speaker 1: how they attract customers, deliver investment services, and maintain control
• 21:04
  Speaker 1: of the investor relationships. For decades, traditional brokers have built
• 21:09
  Speaker 1: their businesses around providing access to financial markets, research tools,
• 21:14
  Speaker 1: portfolio management services, and platforms that allow customers to buy
• 21:19
  Speaker 1: and sell securities. The growth of online trading transformed this
• 21:23
  Speaker 1: model by reducing barriers to entry and allowing individuals to
• 21:28
  Speaker 1: manage investments from their computers and smartphones. AI agents could
• 21:33
  Speaker 1: represent the next major shift moving beyond convenient access to
• 21:37
  Speaker 1: markets and offering investors the possibility of delegating increasingly complex
• 21:42
  Speaker 1: tasks to automated systems. As a result, the competition may
• 21:47
  Speaker 1: no longer be determined solely by which platform offers the
• 21:50
  Speaker 1: lowest fees or the broadest selection of investments. It could
• 21:54
  Speaker 1: increasingly depend on which company can provide the most useful, reliable,
• 21:59
  Speaker 1: and trustworthy AI-powered financial experience. Charles Schwab occupies an important
• 22:05
  Speaker 1: position in this changing environment because its business is built
• 22:09
  Speaker 1: around helping individuals and institutions manage financial assets. Its traditional
• 22:14
  Speaker 1: strengths include brokerage services, investment products, financial research, and relationships
• 22:20
  Speaker 1: with customers who rely on the firm to maintain access
• 22:23
  Speaker 1: to their portfolios. However, the arrival of autonomous AI agents
• 22:28
  Speaker 1: introduces a different kind of customer expectation. Instead of opening
• 22:33
  Speaker 1: a brokerage application to compare securities and manually enter orders,
• 22:38
  Speaker 1: an investor might eventually ask an AI assistant to review
• 22:41
  Speaker 1: a portfolio, identify investments that meet specific criteria, and prepare
• 22:47
  Speaker 1: or execute transactions within authorized limits. This approach could reduce
• 22:52
  Speaker 1: the time spent navigating conventional investment tools, but it could
• 22:56
  Speaker 1: also change the role of the brokerage itself. The platform
• 23:00
  Speaker 1: may increasingly become the financial infrastructure behind an intelligent assistant
• 23:05
  Speaker 1: rather than the primary place where customers make every decision.
• 23:10
  Speaker 1: For established firms, the challenge will be to adopt new
• 23:13
  Speaker 1: technology without weakening the trust and relationships that have supported
• 23:18
  Speaker 1: their businesses for years. Technology-focused trading platforms could have an
• 23:23
  Speaker 1: advantage in attracting customers who are comfortable with automation and
• 23:27
  Speaker 1: expect digital services to operate with minimal friction. Companies such
• 23:32
  Speaker 1: as Robinhood have built recognizable brands around accessible investment platforms,
• 23:37
  Speaker 1: while Coinbase has developed services that connect users with digital
• 23:41
  Speaker 1: assets and related financial products. Their experience operating technology-driven services
• 23:47
  Speaker 1: could help them experiment with new interfaces, automated research, and
• 23:52
  Speaker 1: AI-assisted financial management. Other technology companies could also enter the
• 23:57
  Speaker 1: competition by developing agents that connect to financial accounts and
• 24:02
  Speaker 1: complete tasks across different services. An investor might use one
• 24:06
  Speaker 1: assistant to organize expenses and review savings, compare investment opportunities,
• 24:12
  Speaker 1: and interact with a brokerage account. If such experiences become
• 24:16
  Speaker 1: practical and secure, customers may begin to choose financial services
• 24:21
  Speaker 1: based on how well they connect with their preferred AI
• 24:24
  Speaker 1: tools rather than on the reputation of a single brokerage
• 24:27
  Speaker 1: platform alone. This could introduce new competitors into a market
• 24:32
  Speaker 1: historically dominated by firms with established financial infrastructure. The competition
• 24:37
  Speaker 1: will also involve control over the customer relationship. In traditional investing,
• 24:43
  Speaker 1: brokerage firms generally provide the interface through which investors review
• 24:47
  Speaker 1: their holdings, research assets, and submit transactions. The customer becomes
• 24:53
  Speaker 1: accustomed to the brokerage's application and may use its additional
• 24:57
  Speaker 1: services over time. With AI agents, however, that relationship could
• 25:02
  Speaker 1: become less direct. An investor may ask an independent AI
• 25:06
  Speaker 1: service to recommend an investment, then allow it to communicate
• 25:10
  Speaker 1: with a brokerage to execute the transaction. In that arrangement,
• 25:15
  Speaker 1: the AI provider could become the main point of contact,
• 25:18
  Speaker 1: while the brokerage performs the underlying financial functions. This shift
• 25:23
  Speaker 1: matters because the company that controls the customer interface may
• 25:27
  Speaker 1: have greater influence over which products are considered, which financial
• 25:31
  Speaker 1: services are recommended, and where transactions ultimately take place. Established
• 25:37
  Speaker 1: firms therefore have incentives to develop their own AI capabilities,
• 25:41
  Speaker 1: build partnerships with outside providers, or ensure their systems can
• 25:46
  Speaker 1: safely connect with authorized third-party agents. Fees and operating costs
• 25:52
  Speaker 1: could become another important area of competition. Traditional brokerage services
• 25:57
  Speaker 1: have already experienced pressure from digital platforms that offer lower
• 26:01
  Speaker 1: cost or commission-free trading for many transactions. As artificial intelligence
• 26:07
  Speaker 1: begins to take on a more active role in financial markets,
• 26:11
  Speaker 1: regulators face an increasingly complicated challenge. Determining who should be
• 26:16
  Speaker 1: held responsible when an automated system makes an investment decision
• 26:20
  Speaker 1: that causes financial harm. Traditional investing already involves established relationships
• 26:26
  Speaker 1: between customers, brokerage firms, financial advisors, and the institutions responsible
• 26:32
  Speaker 1: for executing transactions. However, autonomous AI agents introduce additional participants,
• 26:39
  Speaker 1: including software developers, technology providers, and companies that supply the
• 26:44
  Speaker 1: information used to guide automated decisions. When these systems operate together,
• 26:50
  Speaker 1: identifying responsibility for a mistake may become more difficult. An
• 26:54
  Speaker 1: investor could lose money after an agent misinterprets an instruction,
• 26:58
  Speaker 1: a brokerage platform processes an unintended order, or an external
• 27:03
  Speaker 1: service provides inaccurate information. The central regulatory question is how
• 27:08
  Speaker 1: to protect investors while allowing financial technology to advance without
• 27:13
  Speaker 1: leaving customers uncertain about their rights. One of the first
• 27:17
  Speaker 1: issues regulators must address is the distinction between providing financial
• 27:21
  Speaker 1: information and exercising authority over a person's investments. An AI
• 27:27
  Speaker 1: assistant that explains company earnings or summarizes market developments is
• 27:31
  Speaker 1: performing a different function from an agent authorized to buy
• 27:35
  Speaker 1: and sell securities. The consequences of an inaccurate explanation may
• 27:39
  Speaker 1: depend on how the customer uses the information. but an
• 27:43
  Speaker 1: unauthorized or poorly executed transaction can produce an immediate financial loss.
• 27:49
  Speaker 1: As AI systems move from making recommendations to carrying out transactions,
• 27:54
  Speaker 1: regulators may need clearer standards governing the permissions granted to
• 27:58
  Speaker 1: these systems. Investors should understand whether an agent is permitted
• 28:03
  Speaker 1: only to provide suggestions, prepare orders for review, or execute
• 28:07
  Speaker 1: transactions independently within agreed limits. Such distinctions would help establish
• 28:13
  Speaker 1: expectations before something goes wrong and could make it easier
• 28:16
  Speaker 1: to determine whether a system has acted within the authority
• 28:20
  Speaker 1: provided by its user. Accountability becomes especially complicated when several
• 28:25
  Speaker 1: organizations contribute to a single transaction. Consider a situation in
• 28:30
  Speaker 1: which an investor uses an independent AI application connected to
• 28:34
  Speaker 1: a brokerage account. The agent might rely on financial data
• 28:38
  Speaker 1: supplied by one company use a model developed by another,
• 28:42
  Speaker 1: and submit orders through a third-party trading platform. If the
• 28:47
  Speaker 1: system executes an unsuitable trade, the cause might be an
• 28:50
  Speaker 1: incorrect instruction, a software defect, a misleading data feed, or
• 28:56
  Speaker 1: a failure in the brokerage's transaction controls. Each organization could
• 29:01
  Speaker 1: have a different view of its responsibilities, while the investor
• 29:04
  Speaker 1: may have little visibility into the technical process that produced
• 29:08
  Speaker 1: the decision. Regulators will need to consider how responsibility should
• 29:12
  Speaker 1: be allocated when failures involve multiple providers. Clear agreements, reliable
• 29:18
  Speaker 1: transaction records, and well-defined duties could help prevent companies from
• 29:22
  Speaker 1: shifting blame between one another while customers struggle to obtain
• 29:27
  Speaker 1: an explanation or remedy. Another important consideration is the information
• 29:32
  Speaker 1: investors receive before they authorize an AI agent to operate.
• 29:36
  Speaker 1: Customers should be able to understand the system's main functions,
• 29:40
  Speaker 1: its limitations, the types of transactions it can perform, and
• 29:44
  Speaker 1: the risks associated with allowing it to act without direct approval.
• 29:49
  Speaker 1: These disclosures should be understandable rather than buried in lengthy
• 29:53
  Speaker 1: technical documents that few ordinary users are likely to examine.
• 29:58
  Speaker 1: An investor should know whether the agent can trade specific
• 30:01
  Speaker 1: financial products, how much money it can commit, whether it
• 30:05
  Speaker 1: can change an existing strategy, and when human approval is required.
• 30:09
  Speaker 1: The system should also communicate that artificial intelligence cannot guarantee
• 30:14
  Speaker 1: profitable outcomes or accurately predict every market movement. Transparent disclosures
• 30:20
  Speaker 1: would not eliminate investment risk. but they could reduce the
• 30:23
  Speaker 1: likelihood that customers grant permissions based on unrealistic expectations about
• 30:29
  Speaker 1: what automated technology can accomplish. The protection of investors' money
• 30:34
  Speaker 1: will also require rules governing how AI agents obtain and
• 30:38
  Speaker 1: exercise trading authority. A responsible framework could require explicit authorization
• 30:43
  Speaker 1: before an agent begins making transactions, with permissions designed to
• 30:48
  Speaker 1: match the tasks the investor intends to delegate. For example,
• 30:53
  Speaker 1: a customer might allow an agent to rebalance a portfolio
• 30:56
  Speaker 1: within predefined limits while requiring separate approval for investments involving
• 31:01
  Speaker 1: leverage or substantial changes in risk exposure. Systems could also
• 31:06
  Speaker 1: be expected to maintain detailed records showing the instructions they received,
• 31:11
  Speaker 1: the orders they submitted, and the safeguards applied before execution.
• 31:16
  Speaker 1: The growing adoption of artificial intelligence in financial markets could
• 31:20
  Speaker 1: eventually change not only how individual investors manage their money,
• 31:24
  Speaker 1: but also how markets behave as a whole. If AI
• 31:28
  Speaker 1: agents become widely available, millions of people could begin relying
• 31:32
  Speaker 1: on automated systems to research investments, monitor portfolios, and execute
• 31:38
  Speaker 1: transactions according to predetermined instructions. Such a transformation could improve
• 31:43
  Speaker 1: the speed at which information is processed and make sophisticated
• 31:47
  Speaker 1: investment tools accessible to a broader section of the population. However,
• 31:52
  Speaker 1: widespread automation would also introduce a new dynamic into financial
• 31:56
  Speaker 1: markets where the combined decisions of many independent systems could
• 32:00
  Speaker 1: influence prices, trading activity, and investor behavior. What appears beneficial
• 32:06
  Speaker 1: for one individual investor might produce unexpected consequences when thousands
• 32:12
  Speaker 1: of similar agents respond to the same information at approximately
• 32:16
  Speaker 1: the same time. The challenge will be to ensure that
• 32:19
  Speaker 1: the efficiency gained through automation does not come at the
• 32:22
  Speaker 1: expense of market stability. One of the most significant potential
• 32:26
  Speaker 1: benefits is the ability of AI agents to process large
• 32:30
  Speaker 1: amounts of financial information more quickly than many individual investors
• 32:35
  Speaker 1: can manage on their own. Financial markets generate a continuous
• 32:39
  Speaker 1: flow of earnings announcements, economic statistics, interest rate decisions, corporate disclosures,
• 32:46
  Speaker 1: and geopolitical developments. Interpreting these developments requires time and effort,
• 32:52
  Speaker 1: particularly for people who do not work in finance professionally.
• 32:56
  Speaker 1: Automated systems could help organize this information, identify developments relevant
• 33:02
  Speaker 1: to particular investments, and alert users when their portfolios require attention.
• 33:08
  Speaker 1: With appropriate authorization, agents might also carry out routine tasks
• 33:12
  Speaker 1: such as maintaining target allocations or purchasing assets when predetermined
• 33:17
  Speaker 1: conditions are satisfied. This could reduce the amount of time
• 33:21
  Speaker 1: investors spend managing their accounts and make some investment strategies
• 33:25
  Speaker 1: more practical for people with limited resources. If these services
• 33:29
  Speaker 1: become affordable and reliable, Individuals who previously struggled to manage
• 33:35
  Speaker 1: investments could gain access to tools that were once associated
• 33:38
  Speaker 1: mainly with professional financial institutions. Greater accessibility could also influence
• 33:44
  Speaker 1: how people participate in financial markets. Traditional investing often requires
• 33:49
  Speaker 1: a degree of financial knowledge, confidence, and time that not
• 33:53
  Speaker 1: every household possesses. AI agents could help bridge some of
• 33:57
  Speaker 1: these gaps by explaining unfamiliar concepts and translating financial objectives
• 34:02
  Speaker 1: into practical instructions and making it easier to monitor progress
• 34:07
  Speaker 1: toward long-term goals. Investors might receive clearer explanations of portfolio concentration, diversification,
• 34:15
  Speaker 1: and the relationship between risk and potential returns. Automated assistance
• 34:20
  Speaker 1: could also make routine account management less intimidating for newcomers,
• 34:24
  Speaker 1: potentially encouraging broader participation in wealth-building activities. Nevertheless, Accessibility should
• 34:31
  Speaker 1: not be confused with guaranteed financial success. People who gain
• 34:36
  Speaker 1: easier access to trading may still lack an adequate emergency fund,
• 34:40
  Speaker 1: misunderstand the risks associated with particular assets, or invest money
• 34:45
  Speaker 1: they cannot afford to lose. Responsible implementation would require services
• 34:50
  Speaker 1: that communicate uncertainty clearly rather than presenting automated investing as
• 34:55
  Speaker 1: an effortless route to wealth. A more complicated possibility emerges
• 34:59
  Speaker 1: when many AI agents begin responding to the same market developments.
• 35:04
  Speaker 1: Individual systems may use different models and strategies, but they
• 35:08
  Speaker 1: could still react similarly when presented with widely reported information.
• 35:13
  Speaker 1: For example, an unexpectedly strong economic report might prompt numerous
• 35:18
  Speaker 1: agents to increase exposure to certain stocks, while a disappointing
• 35:22
  Speaker 1: announcement could trigger simultaneous selling. If these systems execute transactions rapidly,
• 35:29
  Speaker 1: their combined activity could amplify short-term movements in prices. This
• 35:34
  Speaker 1: effect may be particularly noticeable in securities with limited liquidity,
• 35:38
  Speaker 1: where relatively modest changes in buying and selling pressure can
• 35:42
  Speaker 1: produce disproportionate price movements. The concern is not that every
• 35:47
  Speaker 1: system will behave identically, but that widespread reliance on comparable
• 35:51
  Speaker 1: data Signals or investment objectives could create patterns of synchronized activity.
• 35:57
  Speaker 1: What begins as a collection of individually rational decisions might
• 36:01
  Speaker 1: become a source of instability when those decisions interact across
• 36:05
  Speaker 1: the market. Automated trading could also contribute to feedback loops
• 36:09
  Speaker 1: in which price changes influence the decisions that produce further
• 36:13
  Speaker 1: price changes. Imagine a situation in which several AI agents
• 36:18
  Speaker 1: interpret falling prices as evidence that an investment has become riskier.
• 36:23
  Speaker 1: Their systems may respond by selling, adding downward pressure to
• 36:27
  Speaker 1: the market. The future of investing may be shaped by
• 36:31
  Speaker 1: a fundamental shift in how people interact with financial markets
• 36:35
  Speaker 1: as artificial intelligence moves from providing information to carrying out
• 36:39
  Speaker 1: decisions on behalf of individuals. What once required investors to
• 36:44
  Speaker 1: research companies, compare financial data, monitor market conditions, and manually
• 36:49
  Speaker 1: submit orders could increasingly be handled by AI agents operating
• 36:54
  Speaker 1: under predetermined instructions. This transformation could make investing more accessible,
• 36:59
  Speaker 1: reduce the time required to manage portfolios, and allow individuals
• 37:04
  Speaker 1: to take a more organized approach to their financial goals.
• 37:07
  Speaker 1: Yet the expansion of automation also raises an essential question
• 37:11
  Speaker 1: about the future of financial responsibility. How can investors benefit
• 37:16
  Speaker 1: from increasingly capable machines without losing meaningful control over their
• 37:20
  Speaker 1: own money? For Charles Schwab and the wider financial industry,
• 37:25
  Speaker 1: answering that question will be central to determining whether autonomous
• 37:28
  Speaker 1: investing develops into a dependable financial service or becomes a
• 37:32
  Speaker 1: source of new risks for consumers and markets. One likely
• 37:36
  Speaker 1: outcome is the gradual emergence of a hybrid investment model
• 37:40
  Speaker 1: in which artificial intelligence handles routine tasks while humans retain
• 37:44
  Speaker 1: authority over important financial decisions. Rather than transferring complete control
• 37:50
  Speaker 1: of an account to an automated system, investors may use
• 37:53
  Speaker 1: agents to monitor portfolios, organize research, identify potential opportunities, and
• 38:00
  Speaker 1: maintain investment allocations within clearly defined boundaries. A person saving
• 38:06
  Speaker 1: for retirement, for example, might authorize an agent to track
• 38:10
  Speaker 1: the balance of different investments and recommend adjustments when the
• 38:14
  Speaker 1: portfolio moves away from an agreed strategy. The investor could
• 38:18
  Speaker 1: permit routine changes within a limited range while requiring personal
• 38:22
  Speaker 1: approval for major decisions. This arrangement would preserve many of
• 38:27
  Speaker 1: the practical advantages of automation without assuming that every financial
• 38:31
  Speaker 1: decision should be made independently by software. As the technology improves,
• 38:36
  Speaker 1: the amount of delegated responsibility may increase, but the need
• 38:41
  Speaker 1: to define the relationship between automated action and human judgment
• 38:45
  Speaker 1: will remain important. For individual investors, the value of AI
• 38:50
  Speaker 1: agents may ultimately depend less on their ability to predict
• 38:53
  Speaker 1: the next market movement and more on their capacity to
• 38:57
  Speaker 1: help people make consistent decisions. Financial markets are unpredictable and
• 39:02
  Speaker 1: and even experienced investors cannot guarantee that a particular stock
• 39:07
  Speaker 1: or investment strategy will produce positive returns. Automated systems may
• 39:12
  Speaker 1: nevertheless help users avoid some common mistakes, such as abandoning
• 39:16
  Speaker 1: a carefully considered plan in response to short-term market turbulence
• 39:20
  Speaker 1: or overlooking the need to diversify their holdings. An agent
• 39:24
  Speaker 1: could provide reminders about predetermined objectives, explain how a portfolio
• 39:29
  Speaker 1: has changed, and identify when an investment strategy no longer
• 39:33
  Speaker 1: matches the user's stated preferences. Such assistance could be particularly
• 39:38
  Speaker 1: useful for people who do not have the time or
• 39:40
  Speaker 1: experience to monitor their finances regularly. However, these benefits would
• 39:45
  Speaker 1: depend on whether the agent has accurate information appropriate instructions,
• 39:50
  Speaker 1: and reliable safeguards. Artificial intelligence should support better decision-making rather
• 39:56
  Speaker 1: than create the impression that financial uncertainty can be eliminated
• 40:00
  Speaker 1: through technology. Another important development could be a change in
• 40:04
  Speaker 1: how investors express their financial goals. Conventional brokerage services often
• 40:09
  Speaker 1: require users to navigate menus, examine charts, select securities, and
• 40:15
  Speaker 1: manually configure transactions. AI-powered interfaces could make the process more conversational,
• 40:21
  Speaker 1: allowing people to describe their objectives in ordinary language and
• 40:25
  Speaker 1: receive explanations of the available options. A customer might ask
• 40:30
  Speaker 1: an assistant to review the balance between different asset classes,
• 40:34
  Speaker 1: explain the risks associated with a potential purchase, or identify
• 40:38
  Speaker 1: investments that meet specified criteria. The system could then translate
• 40:43
  Speaker 1: those instructions into practical steps subject to the customer's permissions
• 40:48
  Speaker 1: and the brokerage's controls. This could make financial services easier
• 40:52
  Speaker 1: to understand for newcomers and reduce some of the complexity
• 40:56
  Speaker 1: that discourages people from participating in investment markets. At the
• 41:00
  Speaker 1: same time, simplicity must not come at the expense of
• 41:04
  Speaker 1: informed consent. When decisions involve personal savings, users should understand
• 41:10
  Speaker 1: what they are authorizing, what risks remain, and how they
• 41:14
  Speaker 1: can intervene if the system's behavior does not match their expectations.
• 41:19
  Speaker 1: The responsibilities of established financial institutions will also evolve as
• 41:23
  Speaker 1: AI agents become more capable. Brokerage firms may increasingly compete
• 41:28
  Speaker 1: to provide secure connections between customer accounts and authorized automated systems,
• 41:34
  Speaker 1: while developing their own assistants to help investors manage portfolios
• 41:38
  Speaker 1: and understand financial information.