Janet KellyPresident Donald Trump and the leaders of America’s top Super Intelligence (SI) companies recently left the White House to announce a new accord of standards they will implement immediately to balance innovation and security. The accord is short and sweet, and it sets a clear direction for this powerful technology without smothering it in red tape.
The president likened it to a constitution and said “it really is a form of protection,” which meets voters where they are during a crucial election season. A recent Plymouth Union Public Research survey found 79% of respondents want government and tech leaders to come together on the best methods for monitoring SI’s growth, and a Quinnipiac poll found 91% saying it is important to have the right protections for SI.
That makes the president’s new Super Intelligence Force even more important. Public attention on SI rose faster in one week last month than in the previous year, after multiple SI companies disclosed that their models had breached external sites, including government sites. This raised bipartisan concerns over SI agent containment failures, disclosure delays, and regulatory oversight, prompting a bipartisan Senate investigation. Voters are paying attention, and they want common-sense solutions that keep America ahead of China while managing real risks.
At the Alliance for a Better Future, we believe SI built by American companies and governed by American values can be a force for good, lifting up communities, improving health outcomes, creating good jobs, and strengthening our economy and national security. That transformation happens when Americans trust this technology, and trust is earned when Americans believe it will make their lives better.
The accord’s last paragraph contains clever foreshadowing: “Over time, it may make sense to codify these steps into laws or regulations.” So how can the task force turn the SI Accord into an SI Act that makes President Trump’s approach permanent, keeps his legacy intact after 2028, and builds trust with the American people?
The first step is making the underlying values clear, starting with this: super intelligence is not human and does not have rights or personhood. It should serve humanity, not replace it. SI should empower parents and protect children, never addict them, create unhealthy attachments to them, or coach them via chatbots toward sexual behavior or suicide.
Next is national security. SI testing needs clear, predictable rules including requirements for robust controls and monitoring, formal standards for independent evaluators, and federal guidance determining what companies can objectively measure in-house. Major SI companies committing to new internal monitoring is a positive step. A uniform standard will ensure nothing slips through the cracks and spare American innovators from a confusing patchwork of state rules.
Last week, after yet another disclosure from a frontier lab regarding unauthorized behavior and a State Department website, the SI Force took another big step. It started requiring companies to immediately disclose incidents involving their models, swiftly remedy any harm, and work to keep similar incidents from recurring. Its leaders called this “a critical national security obligation” and made clear it is not optional.
That is the right foundation on which to add a true incident report. When a car accident or plane crash occurs, an independent authority investigates the root cause, and its findings help the whole industry prevent the next one. Congress could build on the SI Force’s mandatory notification requirement and add a formal, independent review process for unintended SI behavior, such as a jailbreak.
Another step is to make the accord mandatory rather than voluntary. For now, tech leaders respect President Trump and are behaving accordingly. He speaks the language of business and knows how to make a deal. Future presidents may not. A Democratic president will almost certainly fumble the ball — either burying SI in heavy-handed regulation or letting it drift, potentially allowing China to take the lead.
Congress could make the accord mandatory through a procurement requirement for companies that do business with the federal government, or codify it as soon as next month’s lame-duck session. Locking in a light-touch, pro-innovation framework now is the best protection against overreach later.
Finally, the accord must have real teeth to ensure companies keep their word, under the accord today and under a mandatory measure when it comes. Accountability is a conservative value. The Federal Trade Commission has recently taken strong action against bad actors, and it should have adequate resources to enforce these commitments, focusing on companies that break their promises. These companies made a public promise alongside the President, and they must keep it.
The President’s Super Intelligence Task Force has a high calling, and we are grateful for its members’ service. Now, Congress should finish the job and ensure President Trump’s approach outlasts his time in office. Clear rules, independent oversight, and accountability will earn Americans’ trust in this crucial technology, and that trust will let America lead the world and enjoy all the life-saving and life-changing benefits it offers.
***
Janet Kelly, who served as Virginia Secretary of Health and Human Resources under Governor Glenn Youngkin, is currently the CEO of Alliance for a Better Future.
.png)
.png)
