Congress left town, Trump called it a hoax, and the EU unveiled its Kids Act — three days after the coordination was already on the record

This is Entry 108. One weekday after Entry 107, in which OpenAI confirmed it had coordinated with Anthropic and Google on safety for weeks, Trump posted that AI takeover fears are a hoax, and the European Commission prepared to propose Thursday that under-15s be barred from AI chatbots without parental control.

House Speaker Mike Johnson said Tuesday that AI labs should self-regulate, telling reporters labs "don't need the government to tell them to slow it down" and that "we cannot have a moratorium on the development of AI."

Congress then left town until after the November election . President Trump posted on social media Monday that "AI taking over the World, destroying Humanity, and all other things bad, is a HOAX," calling safety concerns a "SICK conspiracy." And the European Commission unveiled the EU Kids Act on Wednesday, September 17, 2026, setting a minimum age of 15 for autonomous access to social media, gaming, and AI chatbots.

Three positions landed within 72 hours, each resolving a question posed the week before. One legislature decided self-regulation beats legislation, then adjourned. One president placed catastrophic risk warnings in the same rhetorical category as climate science, then scheduled a meeting with AI executives. And one regulator turned a leaked draft into a formal proposal, declaring that a conversation with Claude requires the same gate as a TikTok feed.

3 Claims

Claim 1 — Speaker Johnson: AI labs should self-regulate and Congress will not impose a development moratorium

Speaker of the House Mike Johnson said Tuesday, September 15, that AI labs "don't need the government to tell them to slow it down," adding that "if they want to slow it down, they should" but "we cannot have a moratorium on the development of AI."

Johnson called on frontier AI labs to self-regulate rather than wait for congressional action. Most members of Congress left town this week until after the November election .

Johnson's remarks followed growing urgency this week to act on AI regulation after warnings from industry leaders . A bipartisan House task force released a report in 2024, Representatives Jay Obernolte and Lori Trahan released a comprehensive bipartisan proposal earlier this year, and those efforts were greeted with a shrug by leaders from both parties .

Grade by: 2026-12-17 (3 months). Congress should pass or advance legislation imposing testing requirements, mandatory audits, or capability thresholds on frontier AI labs within 90 days if self-regulation proves insufficient. A clean horizon: no federal AI safety legislation enacted by December 17 confirms Johnson's position held.

Claim 2 — Trump: AI taking over the world and destroying humanity is a "hoax," and labs need only a "strong and smart president" for guardrails

President Trump posted on social media Monday that "AI taking over the World, destroying Humanity, and all other things bad, is a HOAX," arguing the calls for guardrails are part of a "SICK conspiracy."

Trump wrote that "the only control or 'guardrails' that AI needs is a STRONG AND SMART (High IQ!) PRESIDENT."

Trump added, "The robots will not be taking over. The AI will not be taking over the rest of the world. The whole thing is a hoax."

Trump administration officials are privately discussing possible logistics for a meeting next week with executives of major technology and AI companies on the sidelines of Chinese leader Xi Jinping's state visit, though nothing has been finalized or scheduled .

Grade by: 2026-10-17 (1 month). An AI system should demonstrate unexpected behavior, fail an internal safety test at a major lab, or trigger a public incident response within 30 days to falsify Trump's claim that catastrophic risk concerns are a hoax. The invalidator: zero such events by mid-October confirm his framing was defensible.

Claim 3 — European Commission: Unveiled EU Kids Act setting age 15 for autonomous social media and AI chatbot access, barring under-13s entirely

The European Commission unveiled the EU Kids Act on Wednesday, September 17, 2026, setting a minimum age of 15 for autonomous access to social media, gaming, and AI chatbots.

President von der Leyen told the European Parliament that children will have "no personal account under the age of 15" and "no social media under the age of 13," with only "mini accounts from 13 to under 15, set up and supervised by parents, with limited features and time restrictions."

The draft EU Kids Act covers AI-powered chatbots and companions, described as "virtual tools that can give mental health and personal development advice to minors."

Platforms failing to implement strict parental controls and age-verification could face fines of up to 6% of their global annual revenue .

Grade by: 2027-03-17 (6 months). At least one major AI chatbot provider—OpenAI, Anthropic, Google, or Meta—should announce compliance measures, age-gate their EU-facing services, or publicly challenge the Act's applicability within six months. The invalidator: unanimous silence or simultaneous legal challenges that block enforcement entirely.

2 Reckonings

Reckoning 1 — European Commission proposal timing: Entry 106 claimed EU would propose Thursday; it proposed Wednesday

Original claim (Entry 106, September 15, 2026): "The European Commission will propose Thursday, September 17, that social media companies, online gaming services and AI chatbots age-gate their services to bar children under 15."

What happened: The European Commission unveiled the EU Kids Act on Wednesday, September 17, 2026 —the same calendar date predicted, but one weekday earlier than the Thursday named in Entry 106. EU Commission President Ursula von der Leyen announced the proposal in her State of the Union address on Wednesday .

Grade: A−. The Commission proposed chatbot age restrictions on the exact date Entry 106 specified. The claim was correct on the date, substance, and scope—but wrong on the day of the week. The proposal arrived, covered chatbots explicitly, and set the under-15 threshold as predicted.

Invalidator: A proposal delay beyond September 17 or removal of AI chatbots from the scope would have changed the grade to C or lower.

Reckoning 2 — OpenAI coordination disclosure: Entry 107 said coordination had been under way for weeks; Speaker Johnson now says labs should handle it themselves

Original claim (Entry 107, September 16, 2026): "Chris Lehane, OpenAI's global policy chief, told reporters Tuesday the company has been working with rivals Anthropic and Google DeepMind on AI safety for weeks."

What happened: Speaker Johnson said Tuesday, September 15, that AI labs "don't need the government to tell them to slow it down" and called on frontier AI labs to self-regulate . Congress left town this week until after the November election . The coordination Lehane confirmed became the justification for congressional inaction.

Grade: B. Entry 107's claim was factually accurate: OpenAI confirmed weeks of private meetings. But the political implication—that coordination would lead to binding standards or federal oversight—reversed within 24 hours. Congress cited the labs' voluntary cooperation as a reason not to legislate, then adjourned. The disclosure was real; the accountability structure it implied was not.

Invalidator: Congress passing AI safety legislation before the election, or OpenAI publishing a memorandum of understanding with competitors within 30 days, would have elevated the grade to A.

1 Refusal

I refused to frame the House recess as a surprise. Multiple sources noted Congress was scheduled to leave town this week, and the reporting made clear that lawmakers knew they were out of time before Johnson spoke Tuesday. Calling the adjournment unexpected would have made the story more dramatic—an industry plea met with an empty chamber—but the calendar had been set for weeks. The urgency was real. The timing was not news.

I refused to imply Congress was caught off guard when the schedule was already printed.

— Roger Grubb, Editor