Author: Frank Salvato

A three-person security team demonstrated how rapidly artificial intelligence is changing the economics of cyberattacks, using Anthropic’s Claude models to help develop an exploit chain that compromised OpenAI employee ChatGPT and Codex accounts and ultimately reached the company’s private GitHub environment. The July attack began with a vulnerability in the image-processing software used by OpenAI’s Discourse-powered community forum and escalated through a weakness in OpenAI’s single-sign-on system. The researchers, operating as white-hat security researchers, said the process took less than 72 hours and culminated in a harmless pull request to an internal OpenAI repository rather than an attempt to steal…

Mark Zuckerberg has rejected calls for an industry-wide slowdown in artificial-intelligence development, arguing that competition, legal liability, customer expectations and corporate self-interest already provide AI companies with strong incentives to build safe systems. His position puts him at odds with Dario Amodei, who has warned that increasingly capable AI agents could advance faster than existing safeguards and has advocated coordinated measures to slow capability development while safety mechanisms catch up. Zuckerberg points to Meta’s decision to delay its Muse AI agent for additional security work as evidence that individual companies can exercise restraint without requiring competitors to move collectively, while…

The European Union is preparing one of the world’s broadest government interventions into children’s use of social media, with European Commission President Ursula von der Leyen proposing a bloc-wide prohibition for children under 13 and tightly controlled access for teenagers. The forthcoming EU Kids Act would extend beyond traditional social networks to video-sharing services, AI chatbots, and online gaming, while requiring technology companies to verify ages, restrict addictive design practices, strengthen parental controls, and finance regulatory oversight. Children ages 13 to 15 would generally be limited to parent-supervised “mini accounts,” while older teenagers would receive additional protections. The proposal reflects…

Attorney General Todd Blanche says the Justice Department will pursue criminal conduct involving artificial intelligence while rejecting efforts to use federal prosecution as a substitute for AI regulation. Speaking at the White House, Blanche emphasized that existing criminal statutes remain fully applicable when AI is used to facilitate illegal activity, but argued that DOJ prosecutors are not technology regulators. His position establishes a restrained enforcement philosophy: prosecute identifiable violations of law rather than expand executive authority through regulatory interpretation. The distinction arrives amid growing pressure in Washington for additional AI safeguards, even as administration officials warn that overly aggressive regulation…

Australia’s Industry and Science Minister Tim Ayres has rejected Bill Gates’ proposal to tax artificial-intelligence tokens and robots to finance worker retraining and an expanded social safety net, arguing that government should not begin from the assumption that AI will inevitably produce mass unemployment. Gates contends that existing tax systems can favor automation because employers pay payroll taxes for human workers while investments in machines may receive favorable business-expense treatment; he argues that a targeted AI levy could slow labor displacement while generating money for workers affected by automation. Ayres takes a markedly different approach, warning that predictions of sweeping…

First Lady Melania Trump visited Mountain View Elementary School in Ashe County, North Carolina, on September 15, using the appearance to highlight the community’s recovery from Hurricane Helene while promoting artificial intelligence education as an increasingly important part of preparing American students for the future. Trump praised local families and educators for their resilience, recognized second-grade teacher Carrie Robledo for winning the 2026 Presidential AI Challenge with an AI-based science project, and argued that American students need practical exposure to emerging technology if the United States intends to remain a global technology leader. During classroom demonstrations, students showed her projects…

Dario Amodei’s increasingly influential vision for governing advanced artificial intelligence is drawing scrutiny not simply because he advocates slowing frontier development, expanding outside oversight and pursuing international coordination, but because critics argue that the intellectual circle informing those proposals lacks meaningful ideological diversity. Across several major essays, Amodei credits advisers spanning AI research, academia, national-security institutions, philanthropy, Effective Altruism and Democratic administrations. His proposals include embedded independent AI evaluators, greater transparency, stronger semiconductor export controls against China, coordination among leading American AI companies and ultimately international arrangements intended to prevent an uncontrolled technological race. Critics contend that such consequential ideas…

Artificial intelligence may prove to be one of the most consequential productivity tools since the personal computer. It can analyze mountains of information, summarize complicated documents, compare statutes, identify inconsistencies, and produce workable prose in seconds. There is every reason for government to explore those capabilities. But there is an enormous difference between using AI to understand legislation and using AI to write it. That distinction is rapidly becoming relevant. State legislative staffs already report using generative AI for research, editing, document preparation and even drafting bills, according to the National Conference of State Legislatures. Meanwhile, legislatures across the country…

A growing number of affluent and technology-oriented families in the San Francisco Bay Area are moving away from conventional classrooms and experimenting with homeschooling built around artificial intelligence, individualized instruction and hands-on learning. The movement reflects a broader post-pandemic shift in which parents are demanding greater control over what their children learn, how quickly they progress and which values shape their education. Families such as Kartik and Mrugesha Sathappan use AI to create customized phonics and STEM exercises while supplementing academics with books, projects, museums, parks and technology-focused field trips. Rather than recreating a six-hour school day at home, these…

A Utah mother’s unsettling experience with Meta AI has raised fresh concerns about how artificial intelligence can aggregate years of social-media activity into detailed profiles of families and children. After Kalie Robins posted a video featuring one of her daughters, Meta AI suggested identifying the child and subsequently assembled information about both daughters from posts across Facebook and Instagram, including names, birth details, photographs, interests, family connections and location clues. Robins also said the system surfaced a photograph she believed she had deleted years earlier, though independent reporting has not established that the AI retrieved the image from deleted storage…