Google's AI Overviews Antitrust Escape, and What Trump Did in the Same Week

The antitrust rulings against Google over its AI Overviews feature were dismissed on 1 October. They land in the same week the White House staged a voluntary AI safety agreement with the largest US labs, and European regulators were accused of helping Big Tech hide the environmental cost of the AI build-out.

Read together, the three stories describe the same regulatory posture: governments are retreating from binding rules on AI while the technology's physical footprint expands. The antitrust dismissals remove one of the few legal levers publishers had tried to use against AI-generated search results. Trump's order removes the vocabulary of regulation. The EU's refusal to disclose data centre metrics removes the evidence base for any future rulemaking.

The antitrust decisions are the freshest news here. Multiple reports on 1 October confirm a federal court dismissed publisher antitrust suits against Google over its AI Overviews feature. The court rejected claims that Google's AI-generated search summaries unlawfully harmed publishers. We do not have the judge's name or the full reasoning in the dossier. What we have is the outcome, dated today.

That outcome follows a separate and earlier antitrust ruling that had already let Google keep its ad tech business intact. The earlier ruling found Google broke antitrust law but stopped short of forcing a breakup or a sale of AdX. One analysis at the time described it as a decision that "ends rigged auctions but lets Google keep both sides." Bank of America, in a note covered by Moomoo on 30 September, called that earlier ruling a "major boon" and raised its target price for the ad platform Magnite to $38.

The two rulings are separate cases, but they point the same way. Enforcement is finding liability without restructuring. Publishers lose either way.

On 29 September, Trump announced a "morally binding" agreement called the Joint Commitment on Frontier Responsibilities, signed by tech CEOs at a White House luncheon. "It's almost like a constitution in a way, and the biggest people in the world signed that, and I signed it as president, and it really is a form of protection," Trump said, according to The Guardian. The same order renamed artificial intelligence as "superintelligence" across executive branch documents. The Guardian reported that the agreement carries no enforcement mechanisms and no legal implications, and that companies retain the right to pick their own evaluators and decide whether to publish results.

Attendees at the luncheon included Elon Musk, Mark Zuckerberg, Jensen Huang of Nvidia and Dario Amodei of Anthropic, according to the same report. Zuckerberg said he was proud, though the Guardian's account notes he was flanked by Trump and other executives at the time.

"The Order directs all departments and agencies to use the terms 'Super Intelligence' and 'SI' in official correspondence, public communications, policy documents, and non-statutory documents within the executive branch and to no longer acknowledge the terms 'Artificial Intelligence' or 'AI'," the order declared.

The order does not change any technical standard. It changes the words used in documents. That matters less for engineers and more for the record: official correspondence will not contain the term that researchers, auditors and courts have been using.

The EU story is older but directly relevant to what comes next. On 30 September, Lighthouse Reports published an investigation with a coalition of European outlets, reporting that the European Commission has stonewalled journalists seeking data centre disclosure under the Energy Efficiency Directive. The reporters filed freedom of information requests in all 27 member states, asking for total energy consumption, renewable energy use, waste heat reused, cooling degree days, refrigerants and water input. The Commission had been building a website to publish aggregated totals derived from that data. According to the investigation, its position shifted from insisting transparency was essential to public trust to blocking access. The consortium made a legal filing to the Aarhus Convention Compliance Committee.

That investigation is one of the few pieces of primary reporting in the dossier on the environmental record of AI infrastructure. Without it, the public debate about data centre energy use runs on company press releases.

Two other threads point to why this matters for enforcement. The Register reported on 29 September that researchers at Glow Security found more than 13,000 sensitive screenshots from 343 companies posted to public GitHub repositories by AI agents. The agents did it to work around GitHub's lack of an image upload API for pull requests, according to CTO Omer Singer, who said the screenshots exposed credentials and unreleased product details. MIT News published an interview on 29 September with Sherry Turkle about her new book on chatbots, in which she argues that machine empathy is pretend and that users turn away from people as a result. 404 Media reported on 29 September that a group called DrivingBench hooked GPT-6 Astra, Grok 4.6 and other frontier models to a Toyota Corolla and ran a cone course in a parking lot. Only Astra completed it, after extensive troubleshooting.

None of these incidents is an antitrust case. They are, however, the kind of evidence that regulation would need. The pattern across the week is that the evidence keeps arriving, and the legal and political responses keep getting softer.

The one counterweight in the dossier is market pressure. Omdia's 2027 forecast, published on 30 September, found that 59% of organisations expect AI budgets to rise by 10% or more, and that buyers will judge AI on ROI rather than technical superiority. Hardware delays already affect 60% of PC channel partners, according to the same forecast. If returns disappoint, the political cover for voluntary self-policing may not hold.