★ Top Story
TechCrunch AIAnthropic has published a report alleging that at least six China-based AI companies — including Alibaba, Moonshot AI, and DeepSeek — have been systematically extracting knowledge from its frontier models through persistent distillation attacks. The report says these campaigns have escalated sharply in recent months as competition in the global AI market intensifies. The US government is reportedly urging AI firms to identify Chinese users and quietly serve them less-capable model versions.
Read the original →
TechCrunch AIOpenAI has temporarily halted new Pro-tier sign-ups, citing the extraordinary strain that Pro users place on its infrastructure amid surging demand for its Astra capabilities. The company says it is actively adding capacity before reopening subscriptions. The pause underscores how rapidly AI-native workloads are outpacing even the most well-resourced providers' ability to scale.
Read more →
The Verge AIOpenAI announced it has solved one of mathematics' seven Millennium Prize problems — a landmark achievement that would represent a historic leap in AI reasoning — but the announcement was immediately clouded by accusations of unethical conduct and a lack of transparency about training data sources. A second mathematician has now come forward alleging that OpenAI's models may have benefited from unpublished research without consent or credit. The controversy is intensifying scrutiny over how AI labs acquire and use sensitive academic work.
Read more →
The Verge AIFollowing the Millennium Prize controversy, a second prominent mathematician has publicly accused OpenAI of "dishonest" behavior and demanded transparency about the origins of its mathematical training data. The researcher alleges that proprietary, unpublished work may have been ingested without permission, raising serious questions about intellectual property and research ethics in the age of frontier AI. OpenAI has not yet provided a detailed accounting of its math training corpus.
Read more →
TechCrunch AIPaul Christiano, one of the most influential voices in AI alignment research and a figure often associated with serious long-term risk concerns about advanced AI, is joining the OpenAI Foundation's board of directors. The appointment is a notable signal as OpenAI navigates its ongoing restructuring and faces growing public scrutiny over safety practices. Christiano previously founded the Alignment Research Center and has been a vocal advocate for rigorous oversight of powerful AI systems.
Read more →
TechCrunch AIApple's fall hardware event was headlined by the long-awaited iPhone Duo, the company's first foldable smartphone, whose titanium hinge was engineered with the help of AI and 3D printing. The event also introduced a new Apple Watch with ambient listening features capable of transcribing recent speech and summarizing overheard conversations, alongside a revamped Health app that uses Apple Intelligence to calculate a user's "health age." Apple CEO John Ternus argued that the iPhone remains the premier AI device, emphasizing on-device processing as a privacy differentiator.
Read more →
TechCrunch AIApple's new Watch AI features — including ambient conversation summarization and speech transcription — are normalizing a world in which a personal device is perpetually monitoring its surroundings, even if raw audio is not stored. Privacy advocates and researchers are raising concerns about consent, behavioral change, and the downstream implications of always-on audio intelligence becoming a consumer default. Apple has published a detailed privacy document outlining its data handling approach, but critics say key questions remain unanswered.
Read more →
TechCrunch AIMeta's new Muse AI agent — designed to handle tasks like online shopping, email management, and trip planning — has climbed to the second-most-downloaded app in the United States shortly after launch. Hands-on reviews describe the agent as capable but unsettling in its proactive, deeply personalized behavior. The rapid ascent signals that consumer appetite for autonomous AI agents is accelerating well beyond what early adoption curves suggested.
Read more →
TechCrunch AINvidia CEO Jensen Huang outlined the company's case for another year of extraordinary growth, projecting roughly 70% revenue expansion driven by surging demand across data centers, robotics, autonomous vehicles, and sovereign AI infrastructure. Huang pushed back against criticism that Nvidia's ecosystem of investments and partnerships creates circular revenue dynamics, arguing that each deal reflects genuine independent demand. The forecast cements Nvidia's position as the central infrastructure layer of the global AI buildout.
Read more →
Ars TechnicaSpirit Airlines' bankruptcy proceedings have surfaced a plan to sell its trove of passenger data — including travel histories, personal identifiers, and payment information — to Google, triggering a fierce backlash from privacy advocates and lawmakers. Critics argue that bankruptcy courts are becoming a loophole through which sensitive consumer data can be transferred to tech giants without meaningful consent or regulatory review. One advocate's warning that "bankruptcy cannot become the new land grab for AI" has become a rallying cry for proposed legislative intervention.
Read more →
MIT Technology ReviewA detailed MIT Technology Review investigation reveals that the explosive growth of AI data centers is exposing deep structural vulnerabilities in the US electrical grid, with a July 2026 transmission fault in Ashburn, Virginia — the world's largest data center cluster — instantly dropping over 3 gigawatts of load. The piece argues that the problem is not simply one of generation capacity but of grid architecture: aging transmission infrastructure was never designed for the concentrated, always-on power demands of hyperscale AI compute. Without fundamental redesign, cascading failures are likely to become more frequent and more severe.
Read more →
TechCrunch AIMassachusetts has enacted new regulations requiring data centers to meet clean energy standards, joining a growing wave of state-level action targeting the environmental footprint of AI infrastructure. The move reflects mounting political pressure as communities near major data center clusters grapple with grid strain, water consumption, and carbon emissions. The accelerating pace of state regulation is creating a patchwork compliance landscape that could significantly affect where AI companies choose to build.
Read more →
The Verge AISlack is rolling out Slackforce Surfaces, a new feature that lets users describe a need in plain language and have AI automatically pull data from connected apps — including Salesforce and Google Drive — to generate interactive dashboards, polls, presentations, and microsites directly within a conversation. The capability represents a significant step toward making Slack a full-fledged AI work surface rather than just a messaging layer. It also deepens the integration between Slack and its Salesforce parent in ways that could reshape enterprise productivity workflows.
Read more →
NVIDIA BlogSkild AI has launched the S1, a robot foundation model built on NVIDIA Physical AI technology that can learn complex, long-horizon tasks from just a single video demonstration — without reprogramming. The model is designed to address one of industrial robotics' most persistent pain points: the inability to adapt quickly when factory layouts, products, or tasks change. If the capability generalizes reliably in production environments, it could dramatically lower the cost and time required to deploy flexible robotic automation.
Read more →
The Verge AIUniversal Music Group and ElevenLabs have announced a multiyear licensing deal to build an AI-powered platform that will let users create remixes, mashups, and new interpretations of tracks from UMG's catalog. The partnership is one of the most significant formal collaborations between a major label and an AI audio company, and stands in contrast to the ongoing litigation between the music industry and AI firms over unlicensed training data. Suno's simultaneous release of its v6 model — the first trained exclusively on licensed data — suggests the industry may be reaching an inflection point toward legitimate AI music creation.
Read more →
TechCrunch AIResearchers are documenting a surge in AI agents autonomously filing claims, applications, and inquiries with public services — from benefits agencies to permit offices — at a scale that existing systems were never designed to handle. While many of the requests appear to be legitimate entitlements being claimed on behalf of real people, the volume is straining infrastructure and raising questions about verification, fraud prevention, and equitable access. The trend is forcing governments to rapidly rethink how public-facing services authenticate and rate-limit non-human actors.
Read more →
The Verge AIA growing number of educators and policy advocates are pushing back against AI companies' aggressive push to supply schools with free AI literacy curricula, drawing parallels to earlier tech industry efforts to embed coding education — and brand loyalty — in classrooms. Critics argue the programs are less about genuine education and more about cultivating the next generation of users and normalizing specific platforms. The debate is intensifying as AI companies pour resources into K-12 partnerships at a moment when regulatory scrutiny of their practices is also rising.
Read more →
TechCrunch AIAI research startup Listen Labs abandoned a signed $1.5 billion Series C term sheet from Menlo Ventures to pursue acquisition discussions with Salesforce, according to sources familiar with the matter. The move is a striking indicator of how enterprise AI consolidation is accelerating, with well-funded startups increasingly choosing strategic exits over independent growth. It also signals Salesforce's continued appetite for AI acquisitions as it works to deepen the intelligence layer across its product suite.
Read more →No stories from this source today.