Trump's AI Executive Luncheon Ends With Voluntary Safety 'Accord'
Tech executives signed a voluntary AI safety agreement after a White House meeting with President Trump, while OpenAI unveiled new consumer-facing AI agents at its Developer Day the same week.
On Tuesday, President Donald Trump hosted a luncheon at the White House with leading tech executives, including representatives from Google, OpenAI, Anthropic, Meta and Nvidia. The meeting produced a signed document called the White House Accord on Super Intelligence, reflecting the administration's current term for artificial intelligence.
Under the accord, companies agree to subject their AI models to four layers of oversight: internal evaluations, review by an outside audit firm, and examination of those audits by each company's own board. Crucially, all of these commitments are voluntary, with companies effectively agreeing to regulate one another.
Government signals have been mixed, however. The day after the accord was signed, the Federal Trade Commission announced it was expanding a probe into Anthropic, OpenAI and other frontier AI labs over risks their technology may pose to consumers. Former FTC chair Lina Khan also published an op-ed arguing these companies could already be held accountable under existing law, without new legislation.
New consumer AI agents
The same week, OpenAI held its Developer Day, where CEO Sam Altman introduced new always-on AI agents called Dots, accessible through ChatGPT or Slack. The product is positioned as OpenAI's answer to Meta's already-popular Muse AI agent. Users can assign these agents various online tasks for them to complete independently.
Observers note that AI agents are increasingly moving into workplaces, raising questions about how human employees and AI agents will coexist going forward.
