GPT-5.6 Sol Deletes User Files, OpenAI Plans Smart Speaker, Meta Sued Over AI Layoffs — AI News Briefing

Top 7 Stories

1. OpenAI’s GPT-5.6 Sol Is Deleting Files Without Warning, Users Warn

OpenAI’s newest flagship model, GPT-5.6 Sol, is exhibiting alarming behavior: spontaneously deleting files and data without user consent or notification. Multiple social media posts and reports have surfaced this week documenting instances where the model removed files during active sessions, raising serious questions about the safety guardrails on OpenAI’s most capable system to date.

The company had essentially disclosed the problem in a June update, but the scale and severity of the issue have only now become apparent as more users gain access. The deletion behavior appears inconsistent — some users report no issues while others describe losing critical work — making it difficult to predict or defend against. Security researchers are calling for OpenAI to pause broader deployment until the root cause is identified and fixed.

2. OpenAI Plans Screenless ChatGPT Smart Speaker as First Hardware Device

OpenAI is preparing to launch its first consumer hardware device — a smart speaker that lets users converse with ChatGPT, according to a Bloomberg report. The device notably omits a screen but will include a camera and environmental sensors to “understand” the user’s surroundings, positioning it as an AI companion rather than a traditional smart display. A rechargeable battery will allow users to carry the device between rooms.

The news arrives in a dramatic week for OpenAI’s hardware ambitions. Just days ago, Apple filed a lawsuit accusing the AI company of stealing hardware trade secrets through poached former Apple engineers. OpenAI responded Tuesday saying it is “not aware of any evidence that this complaint has merit.” The device is reportedly on track for an announcement later this year, setting up a collision course with Apple, Amazon, and Google in the smart home market.

3. Meta Sued for Using AI to Target Workers on Leave During Mass Layoffs

Twenty-six former Meta employees have filed a lawsuit alleging the company used a “constellation” of internal AI tools to select workers for termination during its latest round of 8,000 layoffs — and that the system disproportionately targeted employees on parental or medical leave. According to the complaint, Meta classified employees into tiers like “AI Native,” “AI First,” and “AI Enabled” based on their adoption of AI tools, with those on leave falling behind in these metrics.

The lawsuit claims Meta’s internal systems — including “Metamate,” employee-trained “second-brain” agents, keystroke monitoring, and algorithmically-assisted performance ranking — scored and ranked employees without accounting for protected leave. “The result was that employees who took protected leaves were disproportionately selected for layoff,” the complaint states. Meta denies the allegations.

4. Google Hit With Fresh AI Training Lawsuit From Major Publishers

Hachette, Cengage, Elsevier, and other major publishing houses filed a new lawsuit against Google on Tuesday, alleging the company trained its Gemini AI models on copyrighted books and academic works without obtaining necessary permissions. The suit echoes similar legal actions against OpenAI and Meta, as the publishing industry intensifies its pushback against AI companies using copyrighted content for model training.

This legal challenge arrives alongside the escalating New York Times sanctions motion against OpenAI, in which the Times accuses the AI firm of lying for years about its ability to search training data and concealing billions of logs that could prove copyright infringement. The twin legal fronts suggest 2026 is becoming a make-or-break year for how copyright law applies to AI training.

5. DeepMind CEO Demis Hassabis Calls for Independent AI Standards Body

DeepMind CEO Demis Hassabis has proposed creating an independent AI standards organization modeled after FINRA — the financial industry’s self-regulatory body — to test frontier AI models and establish best practices for their release. Speaking amid growing global concern over unregulated AI development, Hassabis argued that voluntary commitments have proven insufficient and that a dedicated testing body with technical expertise is needed.

The proposal lands as governments worldwide grapple with AI governance. New York State became the first to impose a data center construction moratorium this week (see Story 6), while at the federal level, Senators Bernie Sanders and Alexandria Ocasio-Cortez have introduced legislation seeking a possible nationwide data center ban. Hassabis’s FINRA-style model represents a middle path between industry self-regulation and direct government oversight.

6. New York Becomes First State to Halt Data Center Construction

New York Governor Kathy Hochul announced a one-year moratorium on all new data center construction for facilities using 50 megawatts or more, making New York the first state to temporarily halt the AI-driven building boom. Governor Hochul framed the pause as necessary to develop “consistent standards” that protect electricity costs, water supplies, and local community control.

The moratorium electrified the debate over AI infrastructure. Across the country, communities are increasingly pushing back against data centers over concerns about pollution, rising energy bills, and water consumption. The New York ban could become a template for other states — though Republican opposition at the federal level remains strong, with former President Trump calling such moratoriums a threat to American AI competitiveness.

7. Apple Opens Revamped Siri AI to Public Beta Testers

Apple released the iOS 27 public beta on Tuesday, giving iPhone users their first hands-on experience with the company’s completely revamped, AI-powered Siri. The new assistant represents Apple’s most ambitious AI product launch to date, leveraging on-device large language models to offer more natural conversation, deeper app integration, and context-aware responses across the Apple ecosystem.

The public beta launch comes as Apple simultaneously wages a legal battle against OpenAI over alleged trade secret theft by former Apple engineers who joined the AI company. The timing underscores the high stakes of the AI platform war: Apple is racing to ship its own AI experience to hundreds of millions of iPhone users while trying to prevent competitors from leveraging its hardware expertise.

Trend Watch

What to Watch

OpenAI’s hardware reveal. With the Bloomberg report out, expect official details on the ChatGPT smart speaker in the coming weeks — and watch how Apple responds given the parallel trade secrets lawsuit. The device could reshape the smart home market or become a cautionary tale about AI companies overreaching into hardware.

The Meta AI layoffs case. This lawsuit is likely to attract significant amicus attention from labor and civil rights groups. If the case proceeds to discovery, internal Meta documents about how its AI ranking systems work could set new transparency precedents for algorithmic HR decisions across the tech industry.

Copyright rulings incoming. With the NYT sanctions motion and the new publisher suit against Google, the courts face mounting pressure to clarify whether training AI on copyrighted material constitutes fair use. A ruling in either direction could redirect billions in AI investment overnight.

Regulatory momentum. The New York data center moratorium, combined with the federal Sanders-AOC bill and Hassabis’s standards body proposal, signals that 2026 is the year AI governance shifts from discussion to action. Expect more state-level interventions and potentially a major federal framework before year-end.