Good morning and welcome to the agency.[1] I’m going to do something that I never do. I’m going to take credit for the idea behind these roundtables. I can’t take credit for anything beyond that. All of the credit for the hard work, preparation and planning goes to my colleagues in DERA and Exams and elsewhere at the agency – chief among them Rob Fischer, Jill Henderson and Hane Kim. The staff served as the architects who drew up the vision for today’s panels and brought the idea to fruition. And then the deep knowledge and practical input on AI – the expertise that will hopefully make today a fruitful exercise – that comes from our esteemed panelists and moderators. So, really, I did close to nothing. But I will take credit for the original idea.
And I will tell you why. As many in this room will recall, in July 2023, our Divisions of Investment Management and Trading and Markets proposed a rule surrounding the use of Predictive Data Analytics by Broker Dealers and Investment Advisers. It attempted to address how financial professionals dealt with conflicts of interest in light of certain emerging technologies that they might be using in their interactions with investors.
At the time of the proposal, I asked some questions – chief among them was “is the scope of the proposal appropriate?” I reviewed that comment file soup to nuts, and I think it’s fair to say that many – perhaps many of you sitting in the room here today – felt that the scope of the proposal was not appropriate. It was described as overbroad in its definition of covered technology and perhaps overlapping in much of its application with Reg BI and an investment adviser’s fiduciary duties.
Now, whenever I speak with members of financial services community about AI, I ask them – how do you define AI and how are you using it? There is one constant. No one is on the same page. In many ways, when we speak about these technologies, I think we have a tendency to talk past each other.
So, let’s reset. And let’s answer some very basic but important questions.
- What is artificial intelligence? How do we achieve greater precision in our terminology and definitions?
- How is AI being deployed in the financial services industry today? And how will it be used in the future?
- I ask this with respect to all industry participants – how is it used by broker dealers, investment advisers, issuers, intermediaries, markets, investors and other stakeholders?
- Are the most common uses investor-facing, related to investing or the deployment of capital, or are they in pursuit of back-office efficiencies? Is AI being deployed by companies or by their service providers?
- What governance mechanisms are in place to oversee AI systems – especially those systems that may employ “black box” algorithms, where it’s not clear how inputs are weighed or outputs derived? What do those governance mechanisms look like? And is AI being used at board or governance levels?
- And what oversight is applied to ensure that legal and regulatory responsibilities are met – both fiduciary and professional obligations, as well as the relevant rules governing trading mechanics or back-office functions?
- What disclosures are being made around AI uses and risk, and are they consistent and sufficient?
- Are there areas where investors are left vulnerable by the use of AI – either susceptible to fraud or systematic disadvantage?
- Are there systemic market or volatility risks associated with the use of AI that we need to be focused on?
AI appears to represent a sea-change in technology. It is powerful and persistent. It will drive change. The question is – are we prepared for it? I hope that today’s session is an important step for preparedness at the SEC.
[1] The views I express today are my own and not necessarily those of the Commission, my fellow Commissioners, or the staff of the agency.
