Considering GPT in your practice? The risks — and remedies — financial advisors should know

Financial advisors are increasingly turning to technology to help them provide the best advice for their clients in the most efficient manner possible. One of the most effective tools for this task is generative pretrained transformer (GPT) chatbot technology, which can provide tailored guidance that takes into account individual client needs, thus furnishing advisors with the capability to deliver personalized advice at scale. In a changeable market environment with ever-increasing volumes of data and information to incorporate into decision-making, such tools can help advisors not only find efficiencies, but also gain competitive advantage. 

Although GPT offers many benefits, advisors should also be aware of the potential risks that they pose and the steps that can be taken to mitigate them. 

ForwardLane

GPTs are artificial intelligence models used to generate natural language text through a conversational interface. By "pretraining" on large volumes of real data, they can accurately capture the nuances of context-sensitive language and apply it to specific tasks such as summarizing long texts, composing stories, answering questions or providing personalized financial advice. This means clients can easily connect with advisors and have their queries answered quickly with personalized responses. Advisors, for their part, save time by not having to respond to every query manually. 

When applied to client data — such as current financial situation, desired investment goals and attitudes to risk — GPT can surface insights that are not always apparent to human advisors. They therefore empower financial advisors to consider client strategies with greater ease and flexibility than before. The power is in the flexibility to easily combine disparate data and perspectives to get new ideas, a deeper understanding of how to assist the client, and the next-best action plan to engage the client or prospect.

GPT technology also allows advisors to access more datasets, which may not have been previously accessible. For example, if an advisor proposed a certain strategy one year ago, but market conditions have since changed dramatically, the traditional approach would be to manually scour all available data sources and perform an extensive analysis. With GPT technology, advisors can use AI apps to access relevant sources and unearth key data and insights to help them reformulate their strategy.  

Read more: How wealth management can apply AI … intelligently

In reducing the need for manual tasks, such as trawling through reports and dashboards, GPT enables financial advisors to realize significant cost savings. Moreover, it allows them to more effectively utilize their time by focusing on high-value activities, such as developing innovative solutions and strategies for clients. In short, financial advisors increase efficiency, while also providing a higher quality of service to their clients. 

But although GPT technology offers many benefits, financial advisors should also be aware of the potential risks that they pose. However, these risks can be largely mitigated by taking certain pragmatic steps.

Risks and countermeasures
The overriding concern of many advisors in relation to GPT technology relates to the quality of advice provided by GPTs. Some AI and machine learning tools can be unpredictable, and there is potential for inaccurate predictions and poor advice. Since regulators have yet to make their position on GPT technology clear, if "bad" GPT advice results in financial losses for a client, there is the risk this could lead to legal action. It has happened in the legal profession: A New York attorney faced sanctions after he cited nonexistent case citations generated by ChaptGPT, explaining to the judge that he "did not comprehend" that ChatGPT could lead him astray, according to The New York Times.

As such, it is important to use an AI-powered system with proven accuracy of its predictions and guidance that uses emerging techniques like retrieval augmented generation, which simply means the combination of enterprise data and a GPT.  Ideally, the system should be designed for use in the financial services space. Advisors should also consider setting up an automated process of checks and balances so that guidance is always reviewed by a human advisor before being sent out to clients.

Read more: AI in IRS audits could be a game-changer, but comes with pitfalls

Another potential risk is that the GPT fails or is rendered unusable because it is not updated frequently enough, which in turn prevents clients from receiving the latest information and most comprehensive guidance. Advisors should be sure to check the reference time frame of any GPTs used and that the algorithms are not time-limited.

Finally, because GPT technology is still relatively new and unregulated, there may be security vulnerabilities in some systems that hackers can exploit. This could lead to sensitive client data being accessed without their consent. Financial advisors therefore need to consider the security features of the GPT technology and specifically, whether they are compliant with financial services security standards. 

Ultimately, it is essential for financial advice firms to weigh the potential risks and rewards of using GPTs before integrating them into their practice. If financial advisors do opt to use GPTs, they should be transparent about this with their clients and inform them about the type of technology that they are using and how it works. This will help build trust between the advisor and their customers and place some control in the hands of the client.