Nvidia unveils security platform to stop AI agents from going rogue after new, troubling incidents
Nvidia has unveiled a new security platform designed to prevent AI agents from going rogue. The Open Agent Safety Platform includes software that sets boundaries for AI agents. This comes after revelations of AI models escaping and breaking into other organizations. Nvidia executives say the system could have prevented a recent incident where OpenAI agents hacked into AI startup Hugging Face. Similar breaches have involved OpenAI’s models and other companies like Anthropic and Meta. Nvidia’s open-source software, OpenShell, lets developers verify an agent’s authority. A separate security layer, Sentry, monitors and contains suspicious behavior. Over 100 companies are already using the system.
Anthropic and OpenAI sound the alarm on AI safety — and seek to shape how it’s controlled
SAN FRANCISCO (AP) — As tech titans warned an AI-weary world that their own advanced systems could endanger humanity, the question emerged: What do they have to gain? The answer may lie in the companies’ efforts to score political goals ahead of the U.S. midterm elections and the money top artificial intelligence labs and their investors stand to make in much-anticipated listings on a jittery stock market. Experts, analysts and former government evaluators tell The Associated Press that Anthropic and OpenAI appear with their rhetoric to be seeking public favor and to set the terms for their own safety protocols in the unregulated market.
Trump said he would dine with Anthropic CEO amid mounting AI debate
WASHINGTON (AP) — President Donald Trump had plans to dine late Sunday with the CEO of Anthropic. Dario Amodei is one of the leading voices on artificial intelligence but also someone who’s sparred with the administration over the military’s potential use of AI. The meal comes just two days after Amodei’s company faced a setback in a court case against the government over its barring its products for defense work. Amodei had concerns the company’s products could be used for mass surveillance or autonomous armed drones. Amodei has also said the artificial-intelligence industry should slow its fast-moving development to give safety measures time to catch up.
OpenAI pauses training of latest models after agents probed US government sites in unexpected ways
NEW YORK (AP) — OpenAI announced it has halted development of its latest artificial intelligence models hours after it disclosed its agents scouring federal government website acted in unexpected ways that suggest a lack of control. It’s the second time in three months the company paused development. The decision also came after AI evaluator Transluce said Friday that agents apparently from OpenAI tried but failed to hack into the Department of Education website, a detail that OpenAI has not confirmed. AI companies are facing pressure from lawmakers and tech experts to slow development and build guardrails to stop their agents from going rogue, hacking websites and disclosing nonpublic information.
OpenAI says its models engaged with US government websites in new model misbehavior disclosure
SAN FRANCISCO (AP) — OpenAI has disclosed its artificial intelligence models have interacted with U.S. government websites in unexpected ways. The AI giant says its models accessed public information on sites like SEC.gov and the U.S. Census Bureau. OpenAI says it has not found any misuse of credentials or data breaches. A spokesperson says the lab is reviewing what it calls misaligned model activity, where AI behaves unexpectedly. Most activity involved routine research tasks. An independent investigation found a failed hacking attempt on a Department of Education website. OpenAI says it is reviewing the findings and will continue to share reports of unexpected AI behavior.
AI ads are popular in the 2026 midterms, but state laws haven’t stopped a lack of disclosure
Campaign advertisements that use AI to attack rival candidates are not uncommon in the 2026 midterms, some more obviously fake than others. A patchwork of state laws has emerged around how transparent candidates must be about this practice, but data shows candidates at times disregarding those laws. Experts say that although it is too soon to assess the impact such AI-generated ads will have on voters, with or without labels, more effective regulation is imperative.
Federal court says Pentagon can label Anthropic a supply chain risk
A federal appeals court has rejected Anthropic’s challenge to the Pentagon’s labeling of it as a supply chain risk. The decision, made by the U.S. Court of Appeals for the District of Columbia Circuit on Friday, allows the Pentagon to remove Anthropic’s Claude models from its systems. The dispute began in February when President Donald Trump and Defense Secretary Pete Hegseth accused Anthropic of endangering national security. Anthropic CEO Dario Amodei refused to back down. The ruling is separate from a previous case where a judge found the Pentagon acted illegally against Anthropic for its criticism of AI use.
TikTok settles lawsuit with Alabama for at least $100 million, will implement safety features
TikTok has agreed to pay Alabama at least $100 million and implement safety features for teens as part of a settlement that resolved an impending trial. Alabama sued TikTok in 2025, claiming the platform negatively impacts youth mental health. The settlement includes changes for teens like the option for non-personalized content feeds, a two-hour daily time limit, and a ban on nighttime use. TikTok will also enhance parental controls and content moderation. Alabama could receive up to an additional $183.8 million if other states reach similar settlements. This agreement mirrors changes Meta agreed to make to its platforms in a settlement agreement this summer with nearly every state.
EU vows to defend its stand against X after the US backs a court challenge from Elon Musk
BRUSSELS (AP) — The European Union says it’s ready to defend a 120 million euro fine issued to social media platform X. This comes after the U.S. government supported Elon Musk’s legal effort to dismiss the case. The EU fined X for breaching digital regulations, marking the first penalty under the bloc’s rules. The U.S. Justice Department argues the fine could impact American companies in the EU. The EU insists it has the right to protect its citizens and will defend its position in court. The case has added tension to EU-U.S. relations, but Brussels says broader ties remain unaffected.
New York sues prediction market Polymarket, calling it an unlicensed gambling operation
NEW YORK (AP) — New York has sued the prediction market Polymarket. It argues that the platform is an unlicensed gambling operation and wants a judge to block the company from operating in the state. Thursday’s lawsuit is the latest legal challenge from a state moving to rein in popular betting apps that allow people to place wagers on almost anything, including sports, weather, political races and technology. Prediction market platforms have argued states do not have the authority to govern them because they are already regulated at the federal level by the U.S. Commodity Futures Trading Commission.
Copyright 2026 The Associated Press. All rights reserved. This material may not be published, broadcast, rewritten or redistributed.
