What is artificial intelligence—and what are its benefits and risks?

Key Points

    • Artificial intelligence (AI) can take many forms. A lot of what's talked about today is generative AI.
    • There are many business use cases where computers do most of the analysis or content production.
    • There can be risks to using AI at a Registered Investment Advisor (RIA) firm. Safeguards need to be in place to avoid legal consequences.

AI has been in the news a lot lately. Marc Andreessen, a Silicon Valley venture capitalist, believes AI will save the world.1 Others say that AI could cause the extinction of humanity.

Generally, advisors see AI as promising but have taken a cautious approach. Schwab's 2024 Independent Advisor Outlook Study found that slightly more than half (54%) believe AI will have the greatest impact on industry growth over the next three years, but only 23% have implemented AI at their own firm. Four out of five advisors (81%) think that more work needs to be done before AI's full benefits can be realized and 60% say they're not very well equipped to navigate the challenges that AI could bring.3 So, what exactly is AI and what do advisors need to know about the benefits and risks? 

What is AI?

AI refers to a very broad set of capabilities, and there are many misconceptions about what it can and cannot do. It's helpful to first examine where AI technology stands today, as well as its business use cases, to better understand its potential applications in the future.

When people talk about AI, they're typically referring to machine learning (ML) and the information a computer can produce as a result. This is called "generative AI." Generative AI systems create new information based on existing content that's been entered into its database. Basically, a computer is fed a large amount of data and organizes the data itself, looking for patterns that might be useful. Those patterns then allow the computer to respond to prompts with a reconfigured form of that information. 

You may have heard of ChatGPT. It's a chatbot that's been fed an enormous number of transcripts, articles, books, and web pages. All this information allows ChatGPT to answer questions, write basic reports or articles, and even write poems. It also can produce content in specific voices and is getting good at sounding human because it's been trained in how people talk, including celebrities and well-known writers.

At this time, AI is not what's called "artificial general intelligence" (AGI). An AGI system would closely resemble human thought and be able to come up with novel observations and ideas. No AGI system exists yet, but it's not hard to imagine the potential for utopia or disaster if an intelligent computer were in control of major institutions.

Illustration of a robot glitching

How is AI being used in business?

Chatbots are the most common use of generative AI because they can answer questions or produce content quickly. Often, users don't even realize they're talking to a computer when they chat with customer service online. Chatbots have been around for a while, but as AI improves, they're becoming more flexible and better able to serve up complete information. 

Other business applications of AI include:

  • Data analysis—Computers can process data and find patterns much faster than people. With training, AI can make recommendations based on its findings.
  • Automation—Repetitive tasks such as completing forms can be done by well-trained computers.
  • Personalization—AI can make suggestions to users based on known and inferred preferences.
  • Logistics and planning—AI tools can analyze data and help map out more efficient processes.

In a recent Schwab study on AI, some advisors mentioned using AI and ML to create marketing materials, streamline prep for client meetings, and conduct research. One advisor said: "I see AI as a smarter Google search. I sometimes use Open AI (ChatGPT) and get more targeted responses on different topics."4

Even advisors who don't currently use AI are thinking ahead about potential applications. Our 2024 Independent Advisor Outlook Study found that 62% of firms plan to use AI to automate routine tasks. That was by far the most popular use, followed by enhancing risk management and compliance (39%) and marketing (35%).5 It's clear that the first priority for most advisors is to use AI to improve efficiency and free up time for clients.

How Schwab is combining the power of its people and AI

AI can be a useful tool for helping people do their jobs more efficiently—and that's just what Schwab is doing with Schwab Knowledge Assistant, a GenAI tool that provides a ChatGPT-like experience for its client service representatives. The representative simply types in a question and in seconds gets answers in plain language along with links to sources. Hardeep Walia, managing director, head of AI & personalization says, "The combination of cutting-edge AI technology and human insight will be key to delivering the best outcomes for our clients." Explore more about Schwab Knowledge Assistant.

What are the risks of AI for financial advisors?

The more you ask of AI, the more risk is possible. This is because AI still lacks judgment. Consider how a computer works. It stores information in a structured way and then delivers that information when the right commands are given. The current AI does a more sophisticated version of this, but the basic problem is the same: If the prompts are misaligned with the programming, or if the programming itself is faulty in some way, the AI can make mistakes. These mistakes include:

  • Misinformation: Many AI systems will create what you want, even if what you're asking for is not based on reality.
  • Bias: It's rare that AI-generated content is overtly offensive, but assumptions get coded in and can lead to racist, sexist, and other types of biased content being produced.
  • Copyright infringement: AI is trained on existing materials. If it borrows too heavily from those materials and you use what AI created, you could be inadvertently violating the copyright of the original content. Bigger picture, it's still a legal gray area whether tech companies are allowed to input copyrighted content for generative AI.
  • Information leakage: Many AI systems use their prompts as learning input. If your prompt includes sensitive information, it may get repurposed in some way.
  • Relationship damage: If you're using a chatbot to interact with clients, you don't control what it's going to say. It could provide wrong information or say something offensive, which could damage your relationship.
  • Bland content: A publicly available AI tool like ChatGPT is going to produce information that's generally known and believed. It's not likely to offer unique knowledge. It also doesn't know your brand voice or how to talk to your ideal client. The result is often middle-of-the-road content that doesn't differentiate your firm. 

AI is also being used to commit fraud against clients and RIA firms. "I think an area where the bad guys have been using AI to get around controls is the use of deepfake technology to mimic somebody's public voice," says Bashar Abouseido, managing director and chief information security officer at Charles Schwab.6 It's important to be aware of how criminals might use AI to attack your firm so that you're able to prevent losses.

Firms are "dipping carefully" into AI, says Alison Dooher, managing director at Charles Schwab. "We like the cautious approach given the important role that [financial advice] plays in clients' lives."7

All these risks have potential legal consequences, so any firm using AI would need to implement safeguards to make sure the content an AI system produces is compliant and unlikely to trigger a lawsuit.

More change is coming

What will AI be in two years? Five years? Ten? No one knows for sure, but the computing power available today and in the future means AI is likely to evolve significantly.

RIA firms can prepare for coming changes by getting started on policies that can guide AI use by firm employees. In Schwab's AI study, only 7% of firms reported having written controls for AI/ML.8

At the T3 Technology Summit, Dooher also suggested trying out AI for yourself in a low-stakes way so that you can better understand its potential. "I think it's helpful to take a step back, find a safe place to play with the publicly available tool, and then really narrow in on what part of your client experience you're looking to improve."

 

What you can do next

(0425-GK0W)

1Mark Andreessen, "Why AI Will Save the World," Andreessen Horowitz website, June 6, 2023, https://a16z.com/ai-will-save-the-world/.

2Kevin Roose, "AI Poses 'Risk of Extinction,' Industry Leaders Warn," New York Times, May 30, 2023, https://www.nytimes.com/2023/05/30/technology/ai-threat-warning.html

3Independent Advisor Outlook Study, Schwab Advisor Services, October 2024, https://advisorservices.schwab.com/resource/schwab-independent-advisor-outlook-study-2024

4Schwab Advisor Services, "Artificial Intelligence/Machine Learning Research," August 2023.

5Independent Advisor Outlook Study, Schwab Advisor Services, October 2024, https://advisorservices.schwab.com/resource/schwab-independent-advisor-outlook-study-2024

6Schwab Advisor Services, "Unlimited Possibilities, Unparalleled Threats," Webcast, March 2024.

7Sean Allocca, "Financial Advisors Still Cautious about AI Tools," Financial Advisor, January 30, 2024, https://www.fa-mag.com/news/financial-advisors-still-cautious-about-ai-tools-76314.html

8Schwab Advisor Services, "Artificial Intelligence/Machine Learning Research," August 2023.

About the Schwab Advisor Services Independent Advisor Outlook Study
This report was produced by Charles Schwab & Co., Inc. and is intended for independent investment advisory firms. The report is intended for informational and educational purposes only, and is not intended to provide financial, investment, regulatory, compliance, legal, or tax advice. Any guidance taken from the report is not tailored to the particular circumstances of any reader of the report or their firm.

The report relies, in part, on information provided to Schwab by the advisory firms named and others that participated in Schwab's research and interviews. Schwab did not independently verify that information, and Schwab makes no representations about the accuracy of the information in the report. In addition, the experience and practices discussed in the report may not be representative of other firms or the experience or results you might obtain.