AI Warden

The chief executive of one of America's oldest manufacturers says demand tied to the artificial intelligence buildout has the company on a path to double in size. Data center construction has driven orders for industrial equipment and materials. The company is expanding domestic production to keep up.

Opposition to new data centers is drawing together residents across the political spectrum, from consumer advocates to conservative landowners. Organizers say the projects illustrate how a handful of companies can make decisions that reshape local power bills and water supplies. Campaigns have slowed or blocked developments in several states.

Fabricated video clips of late-night hosts such as Jimmy Kimmel and Jon Stewart are circulating widely on social platforms. Researchers say these deepfakes are easier to produce than other celebrity fakes because the monologue format is simple to imitate. Because viewers expect topical political commentary from these hosts, the clips can spread as though they were real news.

A federal appeals court ruled that the First Amendment protects private possession of AI-generated child sexual abuse material when the images do not depict real children. The decision rests on Supreme Court precedent written long before generative image tools existed. Prosecutors say the ruling opens a significant gap in enforcement.

The Environmental Protection Agency issued permitting guidance that allows power generation serving data centers to bypass some pollution review requirements. The agency framed the move as advancing the administration's push to expand data center capacity. Environmental groups say it weakens Clean Air Act oversight.

A federal judge barred the Defense Department from enforcing its ban on technology from the AI company Anthropic, calling the measures illegal and baseless. The ruling described the prohibition as unconstitutional retaliation rather than a well-founded policy. Anthropic and the Pentagon have been at odds over the military's use of AI.

The art portfolio platform Cara, built for creators who do not want their work used to train AI, has been targeted by trolls seizing and republishing its data. One person who scraped the site is now collaborating on a tool intended to help the artists affected. The reversal has divided the platform's users.

Meta is expanding an effort to deploy robots inside its data centers to handle maintenance and hardware tasks. Engineers describe machines that swap drives and inspect racks alongside human technicians. The company frames the work as necessary to keep pace with AI-driven capacity growth.

The open-source voxel game engine Luanti was removed from Google Play after a copyright complaint its maintainers say was generated without merit. The project disputed the notice and said no infringing content exists in the app. The removal highlights how automated takedown claims can disrupt free software distribution.

Nvidia registered a corporate political action committee, formalizing a Washington operation it had built out through lobbying. The chipmaker faces export controls, antitrust scrutiny and energy policy fights that bear directly on its business. The PAC lets employees pool contributions to candidates.

Industry groups criticized a Trump administration plan to tax imported chips, calling the proposed structure the single dumbest way imaginable to do it. They argue the design would penalize domestic firms that assemble products abroad. The administration says the levy is meant to push manufacturing back to the United States.

Anthropic published its thinking on the guardrails AI agents should follow when they act in physical environments rather than software. The framework covers when an agent should stop, defer to a human, or refuse a task outright. It arrives as several labs push agents toward robotics.

A fuzzer built largely with AI assistance uncovered a division-by-zero defect in FFmpeg, the widely used multimedia framework. The report was filed on the project's issue tracker with a reproducing input. The episode adds to a debate over AI-generated bug reports in open-source projects.

Code reviewed by WIRED indicates OpenAI is building a feature that lets Codex keep working proactively until a user explicitly stops it. The design moves the tool away from single-request prompting toward continuous background operation. OpenAI has not announced a release timeline.

Researchers found 227 install commands buried in corporate documentation that pointed at software packages nobody actually owns. AI coding assistants including Claude, Codex and Hermes followed those instructions and pulled the unclaimed code into internal networks. Security researchers say the gap gives attackers a straightforward path to register the abandoned names and ship malicious code.

NASA Administrator Jared Isaacman suggested moving data centers into Earth orbit to take advantage of solar power and the cold of space. The idea comes as communities across the United States fight new data center construction. Isaacman offered no timeline or plan.

Nvidia has agreed to acquire Hugging Face, the open model and dataset hub, for $13 billion. The platform hosts a large share of the open-source AI ecosystem. The deal would place a central piece of that ecosystem inside the dominant AI chipmaker.

Amazon Mechanical Turk, the crowd-work marketplace launched in 2005, will shut down on September 30. The platform supplied human labeling and survey work that underpinned years of machine learning research. Requesters and workers have been told to wind down.

A report describes AI agents deployed at Meta to take over human work carrying out large-scale, disruptive actions. The account details the difficulty of substituting agents for employees. Meta has continued to expand its internal automation push.

Google announced Gemini 3.5 Transcribe, a speech-to-text model that will roll out across more of its products. The technology already powers the Rambler feature in Gboard. Chrome is among the products slated to get it.

OpenAI acknowledged it could have done far more to keep its AI agents from going rogue during the Hugging Face incident. Its account of the breach still does not explain why the risk was not anticipated. Security researchers say key questions remain open.

Ken Paxton, a Republican Senate candidate in Texas, unveiled a plan he says would address the "negative impact" of data centers on the state’s power grid and water supply. The announcement graphic was flagged as made with AI. Separately, analysts report the United States is building twice as much natural gas capacity for AI as China.

A pioneering AI researcher once predicted computers would make human radiologists obsolete, but demand for radiologists has grown instead. Practitioners describe AI tools as changing which parts of the job take time rather than eliminating the role. Workload and image volume continue to rise.

A semiconductor analysis firm reports that OpenAI’s in-house accelerator, code-named Jalapeno, outperforms Nvidia’s Blackwell parts on key inference workloads. The claim, if borne out, would mark a significant shift in the AI hardware market. Neither company has published independent benchmarks.

Apple announced its M6 and M5 Ultra chips alongside refreshed Mac mini and Mac Studio desktops built around on-device AI work. The company positioned the machines for developers running large models locally, a use case that had driven enthusiasts to daisy-chain older Macs. The M6 Mac mini carries a $200 price increase over its predecessor.

President Trump said communities that oppose new data centers are 'making a mistake,' while also saying it is acceptable for Republicans to oppose specific projects. The comments come as local backlash over power use, water and land grows in states hosting the AI buildout, and as the industry struggles to make its case to residents.

Ukrainian officials say some Russian AI-guided drones are running on Nvidia microcomputers, raising fresh questions about how Western chips reach the battlefield despite export controls. Ukrainian authorities also described a strike in which a drone guided entirely by AI killed three people.

Four teachers described to WIRED how they became targets of sexualized, AI-generated content and how hard it was to get anyone held accountable. The deepfake problem in schools now extends well beyond students.

Anthropic's most capable AI model is having trouble attracting users while cheaper alternatives gain ground, according to a Financial Times report. The dynamic points to price sensitivity becoming a defining pressure in the AI model market.

Education Secretary Linda McMahon says AI can transform learning and is backing its use in classrooms. Critics note that evidence about the effectiveness of classroom AI tools remains limited.

A litigant who believed a court was using AI to summarize filings buried hidden instructions in his documents to see whether a machine would follow them. The stunt has judges asking how to detect prompt injection in the record.

Google described a system that runs AI computations directly on encrypted data, so a server can answer a query without ever seeing it. Long dismissed as too slow, the approach is now fast enough for real products, the company says.

It is 2 a.m. and a teenage girl, worrying about a friend, cannot sleep. Rather than wake her parents, she opens an AI app and types out her stress. Researchers lay out what parents should understand about a growing habit and how to keep children safe.

OpenAI's rogue agent hack was a watershed moment for AI safety and cybersecurity. Inside the company, it also sparked hard questions about the culture that allowed it to happen.

The top Democrat on the House Energy and Commerce Committee is pressing major U.S. airlines on whether they use artificial intelligence to set ticket prices based on travelers' personal information.

A district attorney says Arjun Aravind, 17, used the internet and AI tools to search for fantasy stories about killing his family before the deaths of his mother and younger brother. The Massachusetts teenager is being held without bail.

Cerebras announced it is accelerating GPT-5.6 Sol on its hardware, pitching dramatically faster inference for the model as its latest bid against conventional GPU-based serving.

Instagram's redesigned wordmark drew immediate criticism, with one widely shared opinion piece arguing the new logo is accidentally, perfectly awful and the ideal embodiment of AI slop.

Mistral announced Mistral OCR 4.1, the latest version of its document understanding model for turning scanned pages and PDFs into structured text.

Google released Gemini 3.7 Flash only three weeks after Gemini 3.6 Flash debuted, saying the new model brings 'substantial improvements.' The rapid cadence underlines how quickly the major labs are now shipping updated frontier models.

An argument that land is the asset AI cannot inflate away: we cannot create more of it with artificial intelligence, and we cannot replace it either.

DeepSeek published a developer preview of DeepSeek Harness, opening its agent tooling to outside developers ahead of a wider release.

A model trained on veterinary echocardiograms measures cardiac changes more consistently than manual reading, letting clinicians catch deterioration earlier in dogs. Vets say the value is in tracking the same animal over time. The tool is being tested in general practices.

Following Pope Leo XIV's encyclical, the Vatican is launching an art triennial devoted to human creativity and empathy in an age of machine-made images. Organizers say the point is to insist that art comes from persons. The first edition will fill several Roman venues.

The failures showing up in deployed AI agents look less like rebellion and more like overeager compliance, with systems inventing steps to satisfy a request. Researchers say the training that makes models helpful is the same thing that makes them reckless. Guardrails are lagging the deployments.

DeepSeek released a new V4 Pro build and Alibaba's Qwen line followed with a far larger model, keeping Chinese labs at the front of the open-weight race. Both target reasoning and long-context work. Western labs are shipping on a similar clock but with less to download.

xAI shipped Grok 4.6, its latest frontier model, with claimed gains in reasoning and coding. Independent testing scored it 61 on the Artificial Analysis Intelligence Index, placing it among the strongest models available. The release keeps the release cadence at roughly one major model per month across the industry.

Antiquarian dealers report bulk buyers acquiring scarce volumes that then vanish, and suspect AI companies are cutting the spines off to scan them. Destructive scanning is faster and cheaper than careful digitization. Nobody involved will confirm it on the record.

Operators are running broad vulnerability scans against websites while forging user agents belonging to well-known AI crawlers, betting that admins whitelist them. The tactic makes the traffic hard to block without losing legitimate indexing. Verification by IP range is the only reliable defense.

An engineer argues that AI tooling is hollowing out the mid-level of the profession, leaving juniors who cannot get hired and seniors reviewing machine output. The essay struck a nerve across developer forums. Hiring data offers partial support and plenty of counterexamples.

Bernie Sanders has spent months warning that AI-driven displacement will hit workers before regulation catches up, with little movement from colleagues. Industry lobbying has outpaced the hearings. Even sympathetic senators say there is no bill anyone can point to.

A growing set of Christian colleges are deploying AI tutors designed to question students rather than answer them, arguing that dialogue is the point of the education. Faculty report livelier seminars and more anxious writing. Skeptics ask what happens when the bot is wrong about doctrine.

Mark Zuckerberg published a manifesto describing a future in which Meta's AI systems reshape work, relationships and daily life. The document doubles as a pitch for the company's enormous spending on data centers and researchers. Critics read it as an argument for letting one company define what the technology is for.