AIpocalypse.Now
Today'sdoom4.0
Doom 5/10 · 8 stories

Microsoft Named in Times Suit as White House Slows GPT-5.6

Copyright war widens, Washington taps the brakes on OpenAI's next model, and law firms quietly rebuild themselves around the machine.

Published · By · Story-level doom average 3.8/10

The Times Drags Microsoft Into the Pit

The New York Times has amended its copyright complaint and the new version is uglier for Redmond. The paper now alleges Microsoft did not merely benefit from OpenAI scraping its archives, it actively encouraged the practice. That is a meaningful escalation. Up to now Microsoft has played the role of bankroller with clean hands, an investor that wrote checks and shipped Copilot while OpenAI handled the messy ingestion. The amended suit reframes Microsoft as a participant in the alleged infringement, which changes the damages math and the discovery surface considerably.

If the Times can show internal communications where Microsoft pushed for broader training data, the safe-harbor narrative collapses. Every other publisher with a pending claim will copy the language within a week. Microsoft's lawyers know this, which is why the case will probably grind through procedural motions for as long as humanly possible.

Washington Quietly Picks Up the Dial

OpenAI was set to release GPT-5.6. The Trump administration asked them to slow down, citing security concerns, and OpenAI complied by shipping a limited preview instead of a full launch. Read that sentence twice. A US administration that has spent the past year branding itself as the deregulation president just exercised informal pre-release control over a frontier model, and the lab said yes without a public fight.

This is not the EU AI Act. There is no statute, no agency, no defined process. It is a phone call and a delayed launch. Whatever you think of the specific risk in GPT-5.6, the precedent is what matters. Frontier releases are now negotiated with the executive branch, which means future administrations of any flavor inherit a working back channel. The libertarian framing of American AI policy died this week and almost nobody noticed because the press release used the word preview.

It also tells you something about the model. Labs do not delay launches they were proud of. Either the capabilities crossed a line internally, or the red team found something genuinely alarming, or OpenAI wanted political cover for a delay it already needed. None of those readings are reassuring.

Lawyers Discover They Are Both the Problem and the Product

The Financial Times ran a full slate of pieces on the legal industry's awkward dance with AI, and the picture that emerges is coherent if not flattering. Law firms are restructuring, separating casework from back-office functions, building new operational layers around AI tooling. In-house teams insist AI is augmentation, not replacement, eliminating the repetitive grind while preserving headcount. Everyone says this. Everyone has always said this. It is occasionally even true for a while.

The more honest piece in the bundle covers psychology. Legal professionals resist AI for reasons that are not purely economic, including status, craft identity, and a well-earned suspicion of tools that hallucinate citations. Until firms address that, the tech sits unused on expensive licenses. Which connects to the other finding, that despite ballooning legal tech budgets, firms cannot demonstrate returns. You can buy the software. Making partners actually use it is a different project.

The ethics beat is where it gets interesting. Lawyers are not just adopting AI, they are leading the litigation against it, shaping what lawful and responsible deployment looks like in court. The same profession quietly automating its junior associate work is also writing the rules that will constrain every other industry's automation. That is not hypocrisy exactly. It is just leverage, recognized and applied.

The Throughline

Four of today's eight stories are about law, two are about Washington pressuring OpenAI, and one is about the Times suing Microsoft harder. That is not a coincidence of the news cycle. The center of gravity in AI has shifted from capability announcements to questions about who gets sued, who gets to release what, and who restructures around the new tools without admitting it. The engineers built the thing. The lawyers and the executive branch are now deciding what it is allowed to be. Expect more of this, not less, for the rest of the year.

Sources cited in this digest
  1. Times Amends Suit; Microsoft Allegedly Pushed OpenAI to Use Copyrighted Content · New York Times · 6/10
  2. Trump Administration Requests OpenAI Stagger Next Model Release · · 5/10
  3. OpenAI Delays GPT-5.6 Following Trump Administration Security Request · The Verge · 5/10
  4. New Law Firm Structures Emerge as AI Adoption Accelerates · Financial Times · 2/10
  5. Understanding Legal Sector Psychology Key to Overcoming AI Resistance · Financial Times · 1/10
  6. In-House Legal Teams View AI as Workload Augmentation, Not Replacement · Financial Times · 2/10
  7. Law Firms Struggle to Realize Full Value From AI Investments · Financial Times · 3/10
  8. AI Ethics Disputes Trigger New Wave of Legal Challenges · Financial Times · 6/10
Today's doomWeekly column