What does the White House Accord on Super Intelligence mean for US and UK Boards?

Here is the announcement:

Here is a summary of what it means for US and UKBoards:

The Accord commits signatories to four successive layers of control. 

1.        Robust internal monitoring of model capabilities and alignment across cybersecurity, biosecurity, chemical threats and the risk of unintended system access. 

2.        An empowered internal team charged with ensuring those controls function effectively and that problems are remediated. 

3.        Partnership with an independent external auditor or evaluator who will test whether the machinery is operating as intended. 

4.        The designation of an independent committee of the board that receives the reports and is responsible for seeing that identified issues are fixed. The companies further agree to meet regularly to set shared standards.

Implications for US Boards

Under the Accord and existing American fiduciary standards, these are already clear:

·       Directors now carry an explicit duty to oversee independent external audits of Super Intelligence controls and to act on the findings; failure risks Caremark-style oversight liability and the possible denial of D&O coverage.

·       Boards must maintain an independent committee focused on Super Intelligence safety, elevating the issue to the same institutional standing as the audit or compensation committee.

·       Exposure to securities and FTC claims rises sharply if public statements about safety diverge from internal reality, with the availability of punitive damages amplifying the financial stakes.

Implications for UK Boards

British directors, operating under the Companies Act 2006 and without an equivalent Accord, face a different but no less demanding set of pressures:

·       Section 172 requires Directors to promote the success of the company having regard to long-term consequences and reputation, so Super Intelligence risks must now be demonstrably considered in board deliberations.

·       Section 174 imposes a duty of care, skill and diligence that is evolving towards an expectation of basic algorithmic literacy; ignorance is less of a defence than under the American business-judgement rule.

·       Practical barriers to enforcement remain higher: there is no punitive-damages culture, litigation is more expensive and the threshold for success is elevated, leaving weaker private-enforcement incentives than those that now operate in the United States.

The lesson is uncomfortable: liability regimes that expose directors personally and financially are stronger, faster to implement and more effective than statutes. Given the speed at which AIs are developing and their deployment is being implemented, two things stand out:

1.        Boards need to regularly review the implications of any AI agents their company(ies) are using, as well as any AI-based systems they are deploying.

2.         Boards need at least one Director who has a thorough knowledge of and responsibility for existing and future AI capabilities.

3.        Liability insurances need careful reading and consideration before renewal.

If you would like a personal, confidential discussion on this subject, please contact our CEO’s office.