Daily AI brief · October 5, 2026
Trump Launches a Super Intelligence Force
The White House names a federal AI task force led by the DNI and FTC chair, an OpenAI safety lead quits saying its culture is broken, Google Skills start rolling out, and New York City debates AI rules.
Trump launches a "Super Intelligence Force" to run federal AI policy (The Washington Post, TechCrunch, ABC News) On Sunday the President announced a task force to coordinate the federal government's work on AI, which the administration now calls "super intelligence." It is led by Director of National Intelligence Jay Clayton, FTC Chairman Andrew Ferguson, Under Secretary of Defense Emil Michael and Office of Personnel Management Director Scott Kupor, and it reports to the President and chief of staff Susie Wiles. Its charter reportedly tells it to plan responses to AI-enabled threats while "preventing overregulation." Nothing changes in your office this week, but the FTC chair sitting on it is worth noticing: the FTC is the agency that polices how businesses advertise and use AI with customers, so federal AI rules will probably arrive through consumer protection, and the honest-marketing basics apply to you as much as to the big labs.
An OpenAI safety lead quits, saying the company's culture is broken (The Atlantic, TechCrunch) David Robinson, who led the safety reports that ship with OpenAI's major model releases and spent three and a half years there, explained his resignation in an essay in The Atlantic on Saturday. He argues that OpenAI's habit of releasing first and fixing problems afterward gets riskier as the models get more capable. TechCrunch counts him as the fourth safety staffer to leave in about a week, after three were fired. If you run the office, there is no reason to cancel a ChatGPT plan over this, but it is a good reason not to let one vendor's product become the only way your business can do its work, and to keep a second tool you've already tested.
Google Workspace Skills start rolling out today (Google Workspace Updates, Android Central) Skills, the reusable instructions that replace Gems, start reaching Workspace domains on the Rapid Release track today and finish by October 12; Scheduled Release domains start October 19. Gems keep working for now, but business and enterprise accounts lose them on March 1, 2027. The practical part: if your team built Gems for proposals, customer replies or meeting notes, list them this week and decide which ones are worth rebuilding as Skills before the deadline sneaks up.
New York City's whole council debates AI rules today (New York City Council) All 51 members meet today to hear a package of AI bills, including a bounty for AI whistleblowers, a right to sue AI companies for foreseeable harms, and Speaker Julie Menin's bill that would make it unlawful for any business to market, sell or deploy an AI system in the city unless it has passed third-party validation. The heads of Anthropic, OpenAI, Google, Meta and SpaceXAI were asked to attend. These are proposals, not law. Whoever handles AI for you should still watch the validation bill closely, because it is written to cover the businesses that deploy AI, not just the companies that build it.
Move one Gem to a Skill. If you're on Google Workspace, have your AI champion pick the Gem the team uses most, rebuild it as a Skill once it shows up, and run the same three requests through both. If the answers match, you have your template for the rest. If you're on Rapid Release, that can happen this week.
Write down your second option. Take five minutes with the team and name, for each AI task you rely on, which other tool you'd use if your main one changed or went down tomorrow. Write it on the same page as your staff AI rules. If nobody can name one, test an alternative on one real task this month.
Be wary of anyone selling "federal AI compliance." A new task force with a big name will bring a wave of pitches. There is no new federal rule for small businesses today, so don't pay for a compliance package against rules that haven't been written. Truthful marketing claims and clear customer disclosures cover most of what regulators already ask.
Where it's going: Rules for AI are now being written at the city, state and federal level at the same time, and the ones most likely to reach your business will be about how you describe and use AI with customers, not about how the models are built.