To force companies to be more transparent about rapid developments, Illinois would likely rely on “the Big Four accounting and auditing firms—Deloitte, EY, KPMG, and PwC—to audit their safety practices,” Wisor said. The required independent audits will likely frustrate Trump, who has tried and failed to stop states from implementing AI safety laws as Congress stalls on passing any legislation. For Trump, the priority has been to promote AI industry interests, but he began considering expanding federal government safety testing after Anthropic’s Mythos was released and the AI firm limited acces
Trump loses more control over AI regulation as Illinois passes landmark lawUnder the order, Trump directed the National Security Agency to set up a classified benchmarking process to determine the threshold for designating an AI model as a “covered frontier model.” The NSA must also collaborate with the US Treasury Department and the Cybersecurity and Infrastructure Security Agency (CISA) to establish a “cybersecurity clearinghouse” to scan and patch vulnerabilities at scale, as well as a voluntary framework for AI developers to submit models for safety testing. Critics have pointed out, however, that the text of the EO makes it clear how unprepared the government is
Trump plan to test AI models has a problem—US security teams were gutted by DOGE… SpaceX and xAI did not immediately respond to WIRED’s request for comment. The letter also outlines examples of how xAI has not kept up with industry-standard safety practices such as publishing detailed frameworks for mitigating risks around its AI models being used in cyber attacks. …
… Mary Edly-Allen, another co-sponsor, said her goal was to put “minimal guardrails in place” to ensure AI can be a “powerful tool for good.” By passing the law, Edly-Allen suggested that Illinois was creating a “roadmap for responsible innovation to prevent catastrophic risks” rather than an obstacl… …
… In a blog , Microsoft said that “testing for national security and large-scale public safety risks necessarily must be a collaborative endeavor with governments, while crediting the expertise “uniquely held by institutions like CAISI” to conduct such testing. xAI, which is currently fighting agains… …
… A model that appears unremarkable in isolated testing could become a potent cyber tool when integrated into an autonomous pipeline with access to real-world digital infrastructure.” According to Nguyen, the government must be cautious when deciding which models require safety testing, since it risk… …
Hi HN, I built Agent OS because I was tired of the "orchestration tax" – writing the same safety checks, memory management, and tool-handling code in every AI agent project. What it does: - Visual policy edit…
I co-founded a successful security company close to the Mythos ecosystem and have spoken with participants in the know and I am deeply concerned. We, collectively, have answers for some but not all of the problems ahead …
… Trump’s next moves on AI safety in the US won’t just be closely watched by officials within his administration concerned about national security risks—which apparently includes Vice President JD Vance, who said on Wednesday that the administration was prioritizing “protecting people’s data” and “pe… …
Today, we’re publishing the third iteration of our Frontier Safety Framework FSF — our most comprehensive approach yet to identifying and mitigating severe risks from advanced AI models. …
AI + ML Claude Code bypasses safety rule if given too many commands A hard-coded limit on deny rules drops automatic enforcement for concatenated commands UPDATED Claude Code will ignore its deny rules, used to block risky actions, if burdened with a sufficiently long chain of subcommands. …
… "Liability already exists under common law and provides a powerful incentive for AI companies to take reasonable steps to prevent foreseeable risks from their AI systems,” says Thomas Woodside, cofounder and senior policy adviser at the Secure AI Project, a nonprofit that has helped develop and adv… …
… GPT-4o, Grok, and Gemini scored at the highest risks and lowest safety, while the newer GPT-5.2 and Claude Opus 4.5 showed the lowest risk and highest safety. But the things each chatbot said, especially as Lee went deeper and deeper into delusion, were sometimes shocking. …
… Even Geoffrey Hinton, often considered the “Godfather of AI,” has warned that many major tech leaders publicly downplay the risks of artificial intelligence, including job displacement, ethical concerns, or long-term safety, despite understanding them internally. …
