Artificial intelligence is revolutionizing the financial advisory landscape by augmenting human advisors with powerful data analysis capabilities. Asset managers and family wealth advisors have traveled a long way from their early experiments with ChatGPT. They’re now beginning to realize the potential of AI to enhance investment decisions, automate operations and deliver personalized client experiences. These technologies enable advisors to spend less time on routine analysis and more time building meaningful client relationships. But AI opportunities come with unique risks, especially when it comes to data privacy and security, as well as regulatory and legal compliance.
Financial advisors handle highly sensitive data regularly and already operate in a highly regulated sector, so they’re primed to approach AI with risk management in mind. But the use of AI introduces heightened risks in a rapidly shifting environment.
Data Protection
Without data, there’s no AI, but data is also AI’s most significant risk vector. Financial advisors regularly handle individual client data, including tax records, financial plans and other identifiable and sensitive information. Generally, sensitive data shouldn’t be used to train or fine-tune models. Data can be anonymized (or “deidentified,” to use a term that arises in privacy law) to make it safe to use in these settings without exposing individuals’ sensitive information. But frequently, financial sector use cases call for using sensitive data as an input into AI tools. This could include chatbots that are offered as client assistants, predictive modeling for a particular client or family office or even drafting tools used to create content for clients.
The use of AI can expose individuals’ sensitive information to heightened risks. Although data breaches are always a risk that’s hard to control, they can be heightened with the use of unsecured application programming interfaces (APIs) and other third-party AI tools. These external connections can be vulnerable entry points for malicious actors that are difficult to solve for because of the necessary reliance on vendors or partners to deploy proper protections. In the context of AI, where models may handle sensitive, proprietary or confidential information, the stakes are increased.
Model risk and transparency. AI models can be opaque. Their internal decision-making processes aren’t easily understandable or transparent, even to experts. If a model or a tool makes a prediction or a recommendation that a financial advisor acts on, that advisor will need to be able to explain the decision, which can be difficult or even impossible. For instance, a predictive model might flag a new or prospective client as “high risk” or recommend a trade that ultimately leads to a loss. Human-driven systems have been making such predictions for decades, but when AI makes them, humans struggle to explain them.
Lack of AI transparency can erode client trust and put advisors in conflict with regulators, who increasingly require transparent and auditable decision-making processes with respect to AI. For instance, when clients receive recommendations (such as investment strategies or risk profiles) without apparent clear reasoning, they may question the credibility or fairness of the advice. And if advisors can’t demonstrate how AI-driven decisions align with their fiduciary duties or compliance standards, they’re likely to trigger scrutiny.
Legal and regulatory development. In addition to model transparency, regulators are racing to catch up with AI developments broadly. The legislative and regulatory landscape around AI is evolving almost as quickly as the technology itself. The Securities and Exchange Commission, Financial Industry Regulatory Authority and other regulatory bodies have issued preliminary guidance on AI; dozens of countries and multiple U.S. states have passed laws regulating AI; privacy laws in the United States and internationally have outsize impact on the adoption of AI tools; and new guidance, regulations and laws emerge almost daily.
Against this landscape, financial advisors need to adopt AI governance and risk management approaches that satisfy regulators and are sufficiently flexible to address additional emerging rules.
Bias and discrimination. The recurring theme that AI is only as good as the underlying data is important. If AI models are trained on biased data, they’ll perpetuate and even amplify those biases. To take a simple example: An advisor might adopt a tool that can be used to predict client risk based on a set of data points that include ZIP codes. But unbeknownst to the advisor (and perhaps even to the developer of the tool), the datasets training the model that the tool was built on included certain data points that identify ZIP codes of certain historically marginalized communities as indicators of financial risk. The result could be a tool that automatically flags any potential client with one of those ZIP codes as high risk and recommends rejecting that potential client.
Reputational and operational risks. Data breaches, model failures or missteps in automation can create public relations crises, particularly for family offices and asset managers with long-standing reputations. Without strong controls to identify issues promptly and address them effectively, problems can spiral rapidly into reputational damage. Potential perceptions of bias or discrimination in AI-powered decision making must be top of mind. Additional priorities are transparent reasoning in guiding clients and prudent data handling.
Another consideration is that overreliance on AI can erode internal experience and expertise over time. Advisors integrating AI tools must be thoughtful about the potential for reducing opportunities for professionals to apply judgment, solve complex problems and build domain-specific knowledge. In highly skilled sectors such as financial management and advising, overreliance on AI can result in a less capable workforce, weakening an organization’s ability to respond effectively when AI tools fail or nuanced decision-making is required.
*This article is an abbreviated summary of “Artificial Intelligence in Asset and Wealth Management,” which appears in the July/August 2025 issue of Trusts & Estates. Continue reading about AI best practices here.
#Opportunities #Risks #Financial #Advisors