The US Securities and Exchange Commission (SEC) has proposed new restrictions for brokerages and money managers utilizing artificial intelligence (AI) to interact with clients.
SEC Chair Gary Gensler aims to address conflicts of interest that may arise when financial firms adopt these technologies.
As part of this effort, the SEC has also finalized rules mandating companies to disclose significant cybersecurity incidents within four business days.
Under the proposed AI regulations, companies must assess whether their use of predictive data analytics or AI results in conflicts of interest.
They are required to eliminate these conflicts and enhance written policies to ensure compliance with the rule.
Gensler emphasized that these rules are designed to protect investors and enforce firms to prioritize clients' interests.
The concern is not just about disclosure but also about identifying whether predictive data analytics favor financial firms or genuinely serve clients.
While banks and brokerages have long employed AI for fraud detection and market surveillance, there's been a recent shift towards AI-driven trading recommendations, asset management, and lending.
The SEC's intention is to prevent companies from prioritizing their interests over those of their clients when offering trades or products.
The proposed regulations go beyond existing requirements that only mandate brokers to act in their clients' best interests when making recommendations.
The plan will be open to public comments before the SEC finalizes it, likely in 2024, subject to a majority vote from the five-member commission.
Two Republican commissioners criticized the proposal, considering it too broad and requiring assessment of various types of technologies for potential conflicts. They fear that this vagueness and compliance burden could stifle innovation.
The SEC's focus on AI oversight comes as other regulators, like the Consumer Financial Protection Bureau and the Federal Reserve, are also stepping up scrutiny on AI usage in finance to ensure fair practices and mitigate biases.
Additionally, the Federal Trade Commission is investigating OpenAI Inc., the creator of ChatGPT, over potential risks to consumers' data and reputations.
President Joe Biden has also expressed his administration's intention to set a framework for "responsible innovation" with AI.
Chair Gensler has previously expressed concerns about AI's potential to leverage vast data to target individual investors and influence their trading and financial decisions, posing risks of market instability.
In addition to the AI regulations, the SEC has approved rules requiring companies to disclose significant cybersecurity breaches.
The final rule mandates the disclosure of breaches within four business days, but an option for delay exists if the US attorney general deems it necessary for public safety or national security.
The SEC is also considering a proposal to allow investment advisers operating solely online to register with the commission, affecting approximately 200 investment advisers under the current exemption.
