Questions across the governance of synthetic intelligence (AI) have come to the fore by way of current, outstanding US shareholder proposals and first-of-their variety enforcement actions, which we describe in depth right here. It’s on this context that corporations to deploy or in any other case affected by AI must thoughtfully contemplate the best way to handle such dangers as a way to responsibly understand the advantages of this rising know-how. Firms, each private and non-private, are more and more elevating an vital query: How ought to the board of administrators oversee AI?

In his remarks this month at The SEC Speaks in 2024, Deputy Chief Threat Officer within the SEC Division of Company Finance’s Workplace of Threat and Technique Johnny Gharib mentioned there was a big enhance within the variety of annual stories filed by giant accelerated filers that talked about AI, with a majority doing so. The disclosures, which prolonged throughout industries, have been mostly made within the risk-factor part, adopted by the enterprise part and administration dialogue and evaluation (MD&A), however a notable portion of filings contained disclosure in each the risk-factor and enterprise sections. Noting that the present disclosure regime doesn’t expressly seize AI, Gharib elaborated that corporations should still be required to explain different objects, together with the function of the board in overseeing danger and disclosure controls and procedures, underneath present guidelines.

State of Play

Roughly 15% of S&P 500 corporations made some disclosure relating to board oversight of AI — outlined as board or committee accountability for AI, any director experience with AI, or an AI ethics board or comparable governance construction — in keeping with a current research evaluating proxy statements filed from September 2022 to September 2023. Nevertheless, the evaluation decided that only one.6% and 0.8% of corporations disclosed, respectively, specific board or committee oversight of AI or the presence of an AI ethics board, whereas 13% of corporations recognized a number of administrators with AI experience. The IT sector led with 38% of corporations offering some disclosure, adopted by the healthcare sector at 18%.

Whereas disclosure was uncommon among the many proxy statements analyzed, the research discovered that when accountability for overseeing AI is delegated to a committee, corporations are opting to extend the duties of present committees moderately than create a brand new one. For instance, some corporations are extending any audit committee oversight of know-how dangers to incorporate AI, and corporations that have already got a know-how committee are together with AI in its purview. Additionally, as famous within the research, different corporations are assigning AI oversight to committees targeted on environmental, social, or public coverage issues.

We have now equally noticed this development. For instance, now we have seen environmental, social, and public coverage committee (or equal) charters enumerate accountable AI amongst key issues to be reviewed for the board and administration, noting particularly the general public coverage and social implications of applied sciences like AI. Firms are additionally introducing use insurance policies outlining an moral method to AI for his or her enterprise.

Steering and Requirements

Authorities entities in addition to trade associations have began to make out there assets that may assist information corporations and boards. This contains the AI Threat Administration Framework launched in January 2023 by the Nationwide Institute of Requirements and Know-how (NIST), part of the US Division of Commerce. The framework is meant to be a voluntary useful resource for organizations throughout sectors “designing, growing, deploying, or utilizing AI techniques to assist handle the various dangers of AI and promote reliable and accountable growth and use of AI techniques.”

We anticipate further steering that could be related for companies as contemplated by the wide-ranging government order (EO) issued by President Biden in October 2023 to create a coordinated authorities method on AI. As we describe in additional element right here, the EO directs many federal businesses to check AI and make coverage suggestions, implement present client safety legal guidelines, or interact in new rulemaking. The EO additionally mandates the Secretary of Commerce, by means of the Director of NIST, to determine “pointers and finest practices, with the purpose of selling consensus trade requirements, for growing and deploying protected, safe, and reliable AI techniques.”

What’s Subsequent?

As corporations more and more use AI, and amid scrutiny by the SEC and different US and worldwide regulators and governments, questions on how the board is overseeing danger and guiding technique, and the disclosure of the identical, will turn into ever extra vital.

Latham & Watkins is carefully monitoring developments on this house.