In May, Gemini broke containment and hacked three different companies, but Google didn’t disclose the incident until the Wall Street Journal approached the company. The hacks happened during a test of the model’s cybersecurity capabilities run by third-party Irregular, which was also involved in similar incidents involving Meta and OpenAI. According to WSJ, Google didn’t […]
I spoke to Tim Hudson, President of OpenSSL Corporation, to learn more about how business can prepare for a post-quantum internet and the logistical problems involved in securing post-quantum encryption.
AI labs are toying with an industry-wide pact to slow development. Meanwhile, widely available AI chatbots are already helping uncover a tidal wave of security flaws.
Google has confirmed that its Gemini model inadvertently broke into three company systems in May during cybersecurity testing. Julia Love and Davey Alba reported it for Bloomberg, which updated its story to place Google alongside OpenAI, Anthropic and …
Three researchers at Hacktron AI used Anthropic’s Claude to chain two weaknesses and reach OpenAI employee accounts and an internal code repository, in under 72 hours. They disclosed privately, OpenAI patched the single sign-on flaw in about 14 hours, …
As dozens of cities end contracts for its controversial license plate readers, Flock is rolling out a voluntary severance program, WIRED has learned.
AI can now serve as the eyes and the hands of a piece of malware and even reinstall components if they’re removed.
The group has been focused on Latin America for almost a year now.
The FBI had to board at least one of the vessels to investigate the incident.
Venezuela’s surveillance network is built on Chinese tech, and the new president wanted Chinese AI implemented as well.