> ## Content Index
> Fetch the complete content index at: https://www.consumernews.ai/llms.txt
> Use this file to discover other available public pages before exploring further.

# Europe joins the AI slowdown movement as von der Leyen calls frontier labs to the table
- URL: https://www.consumernews.ai/europe-joins-the-ai-slowdown-movement-as-von-der-leyen-calls-frontier-labs-to-the-table/
- Published: 2026-09-17T14:07:39.000Z
- Updated: 2026-09-17T14:07:39.000Z
- Description: AI is becoming a safety issue, not just a free speech issue.
- Author: James R. Hood
- Tags: AI/Privacy

## 

Europe’s top executive has now embraced the “slow the frontier” argument. In her September 16 State of the Union address, European Commission President Ursula von der Leyen [said](https://cyprus.representation.ec.europa.eu/news/2026-state-union-address-president-von-der-leyen-2026-09-16%5Fen?ref=consumernews.ai) increasingly capable models pose risks from advanced hacking and self-improving systems, cited recent incidents involving autonomous agents, and said that when the CEOs building frontier AI say development needs to be paced, governments should listen. 

More importantly, she announced that she will summon the major frontier AI laboratories for talks and said Europe will work with Canada, Britain and other partners on model evaluation, verification, early-warning systems and AI security. This is not an endorsement of an indefinite moratorium; her formulation is closer to Anthropic CEO Dario Amodei’s proposal to slow capability growth when safety controls cannot keep up. 

That matters because the slowdown argument has now jumped from researchers and Silicon Valley executives to the head of the European Commission — an institution that already has statutory AI-regulatory machinery through the EU AI Act. In other words, the debate is beginning to move from warnings to government coordination.

There is also a noteworthy second development. Google DeepMind co-founder Shane Legg has launched the DeepMind Institute to examine how AGI should be deployed and warned that capabilities must not outrun safety controls. Legg [said](https://www.reuters.com/business/deepmind-co-founder-warns-ai-capabilities-must-not-outrun-safety-controls-ft-2026-09-16/?ref=consumernews.ai) Amodei’s slowdown proposal was “worth considering” and reiterated his forecast of a 50% chance of reaching a minimal form of AGI by 2028\. 

### Other AI safety moves

- **The three leading U.S. frontier labs are now coordinating on AI safety.** OpenAI says it has been in talks for several weeks with Anthropic and Google DeepMind about joint safety efforts, and OpenAI policy chief Chris Lehane said the company would support bipartisan legislation aimed at mitigating catastrophic AI risks. Reuters says the companies are discussing cooperation without seeking an antitrust waiver. That is a notable escalation from individual CEOs issuing warnings to the major labs actually trying to coordinate their behavior.
- **The political coalition for slowing AI is becoming remarkably strange and broad.** Bernie Sanders and Steve Bannon appeared at the same Washington gathering Tuesday and [both argued that AI development needs to be slowed](https://www.reuters.com/legal/legalindustry/us-senator-sanders-podcaster-bannon-longtime-foes-urge-ai-restrictions-2026-09-14/?ref=consumernews.ai). Sanders called for binding safety rules and urged President Trump to negotiate with China over a pause; Bannon said the pace should be slowed immediately and criticized efforts to block state AI laws. At the same time, House Speaker Mike Johnson rejected a moratorium and said companies should self-regulate. That gives the issue a genuine cross-partisan fault line rather than the usual left-versus-right regulatory split.
- **Another former DeepMind safety researcher has gone public with an extinction warning.** [Bilal Chughtai](https://www.consumernews.ai/veterans-turned-out-big-for-trump/), who left Google DeepMind in July after working on AGI safety and alignment, said this week that AI could “kill all humans” and that time may be running out to prevent it. He joins Jacob Coxon and Anthropic researcher Evan Hubinger in publicly attaching very high stakes to the current development race.

There is also a consumer-democracy angle from Nobel Peace Prize laureate Maria Ressa, who argued on September 15 that AI should be treated as a public-safety issue rather than merely a speech or innovation issue. She warned that agentic AI could erode human agency and called for governments to hold technology companies responsible for harms, [Reuters](https://www.reuters.com/legal/litigation/limit-ai-now-or-humanity-will-lose-its-agency-says-nobel-laureate-ressa-2026-09-15/?ref=consumernews.ai) reported.

The strongest ConsumerNews.ai framing may now be **“The AI safety debate is turning into a coalition.”** What started as warnings from a few researchers has, within days, drawn in frontier-lab CEOs, former safety engineers, labor voices, politicians from Bernie Sanders to Steve Bannon, and international figures such as Ressa. The central disagreement is shifting from *whether AI poses serious risks* to *whether governments or the companies themselves should be the ones to slow it down*.