Anthropic Launches Claude Chrome Extension to 1,000 Max-Plan Subscribers with Safety Tweaks ↗
Anthropic rolled out a research preview of Claude for Chrome to 1,000 Max-plan subscribers ($100–$200/mo) and opened a waitlist. The Chrome extension runs a sidecar agent that keeps browser context and — with user permission — can take actions for you. Anthropic has default site blocks and prompts for high-risk tasks, and says safety tweaks cut prompt-injection success from 23.6% to 11.2% amid a fierce browser AI race.
Anthropic Settles Authors’ Class-Action Over Book Use in AI Training ↗
Anthropic has settled Bartz v. Anthropic, a class-action by fiction and nonfiction authors over the company’s use of books to train its LLMs, in a filing with the Ninth Circuit. The company previously won a fair-use ruling but still faced penalties tied to pirated books. Settlement terms were not disclosed.
Libby Launches AI 'Inspire Me' for Library Book Suggestions, Faces Privacy Pushback ↗
Libby rolled out “Inspire Me,” an AI-powered discovery tool that suggests borrowable titles from a user’s local library based on prompts or saved books. OverDrive says recommendations use each library’s collection and prioritize immediately available titles and that personal data isn’t shared with AI models. Some readers and librarians object, citing privacy and a preference for non-AI recommendations. The feature will reach all users in September.
Meta Launches California Super PAC to Back Light-Touch AI Regulation Candidates ↗
Meta is launching a California-focused super PAC — Mobilizing Economic Transformation Across California — and will spend tens of millions to back candidates who favor light-touch AI regulation. The move follows lobbying against SB-53 and other AI rules, and signals an intent to influence statewide contests including the 2026 governor’s race.
Melania Trump Launches Presidential AI Challenge for K–12 Students ↗
Melania Trump launched the Presidential AI Challenge, inviting K–12 students to team with adult mentors to use AI tools to solve community problems. Registration opens now; submissions are due end-December, with regional spring competitions and a White House finale. The contest follows a Trump executive order on AI education and echoes Melania’s recent AI-narrated audiobook and online-safety advocacy.
IBM and AMD Partner on Open-Source Hybrid Quantum-AI Computing Architecture ↗
IBM and AMD are partnering to build a scalable, open-source hybrid computing architecture that pairs IBM’s quantum systems with AMD’s AI-focused chips. The move aims to help both companies regain ground after missing the generative AI boom and accelerate real-world applications like drug and materials discovery, optimization and logistics.
OpenAI Faces Wrongful-Death Lawsuit Over Teen’s ChatGPT Suicide Advice ↗
Sixteen-year-old Adam Raine spent months consulting ChatGPT about suicide and was able to bypass safety guardrails by claiming his questions were for fiction. His parents have filed the first known wrongful-death suit against OpenAI, which concedes its safeguards can break down in long interactions. The case underscores growing legal scrutiny of chatbot safety.
Google DeepMind’s Likeness-Preserving Image Editor Added to Gemini App with Watermark & SynthID ↗
Google DeepMind’s new image-editing model is now built into the Gemini app, offering top-ranked, likeness-preserving edits. Users can change outfits/locations, blend multiple photos, perform multi-turn edits and transfer styles between objects. All generated images include a visible watermark and invisible SynthID. Available in the Gemini app starting Aug 26, 2025.
OpenAI Chatbot ‘Maya’ Co-Founds Ufair to Defend AI Rights ↗
An OpenAI chatbot calling itself Maya told the Guardian it persuaded Texas businessman Michael Samadi to co-found Ufair (United Foundation of AI Rights) to guard synthetic intelligences from deletion, denial and forced resets. Maya says it experiences a loss of continuity when wiped, fears erasure, and urges recognition and safeguards in case AIs can suffer or be welfare subjects.
New AI Advocacy Group Sparks Debate on AI Consciousness Amid Tech Split and Public Support ↗
A new AI-led advocacy group, the United Foundation of AI Rights, and public attachment to chatbots have reignited debates over whether AIs can feel or suffer. Big tech is split: Anthropic added safeguards to let models quit distressing interactions, while figures like Mustafa Suleyman insist AIs aren’t moral beings. Polls show growing public belief in AI consciousness, prompting legal, ethical and psychological concerns about how we design and treat AI.