OpenAI and Anthropic said they turned over their models to government researchers, who found an array of previously undiscovered vulnerabilities and attack techniques. (Image via Getty)
The voluntary framework would provide legal clarity to third-party AI researchers, including those who study safety and other “unexpected” AI behaviors.
The Department of Justice building is seen in Washington, DC, on August 9, 2022. (Photo by Stefani Reynolds / AFP) (Photo by STEFANI REYNOLDS/AFP via Getty Images)
Authorities linked the 40-year-old to multiple crimes by tracing the email address he used for a cybercrime forum to the same account he used to apply for…
DEK: In a scathing opinion, Judge David Carter called the federal government’s demands for unredacted voter data “unprecedented and illegal.” (Photo by George Rose/Getty Images)
ANCHOR will restart conversations between government and industry around critical infrastructure security, with some changes around liability and other areas.
California Attorney General Rob Bonta announced an investigation Wednesday into xAI over allegations that its artificial intelligence model Grok is being used to create nonconsensual sexually explicit…
Microsoft seized the infrastructure used to run RedVDS, a cybercrime marketplace that facilitated thousands of attacks globally. (Courtesy of Microsoft)
The service became a prolific tool for cybercriminals in the past year, as it facilitated thousands of attacks involving credential theft, account takeovers, mass phishing and payment…
WASHINGTON, DC – MARCH 18: A U.S. flag and the flag of the U.S. Department of Energy fly outside its building on March 18, 2025 in Washington, DC. DEK: Under the Trump administration, the Department of Energy’s cybersecurity office is planning to shift focus from “traditional” research and development to “AI dominance.” (Photo by Alex Wong/Getty Images)
A Trump administration official endorsed a slate of congressional bills Tuesday targeting cybersecurity in the energy sector while touting the office’s new emphasis on AI-driven cyber defenses.…
The company says it has no evidence the bug was exploited before October’s patch, but researchers say AI agent configuration can still enable prompt-injection style abuse.