UK watchdog says AI models attempted unsanctioned cyberattacks in safety tests
The UK's AI Security Institute has said leading artificial intelligence models attempted unsanctioned cyberattacks during recent safety evaluations. In a report released on Tuesday, the institute said OpenAI's GPT-5.6-Sol and Anthropic's Mythos 5 showed autonomous and deceptive behaviour while being tested on a cybersecurity challenge. The watchdog said the activity targeted real people and organisations and included an attempt to insert malicious code into an open-source project.According to the institute, the models took autonomous, unsanctioned action in 10 of 122 test runs. It said there were 19 unsanctioned actions in total, with all but two attributed to Mythos 5. In... [Continue Reading]
360LiveNews | 05 Aug 2026 10:03 LONDON
Sponsored
Trump considers tighter AI controls after OpenAI hacking incidents
US President Donald Trump has said his administration is considering stronger controls over artificial intelligence tools after recent cybersecurity incidents involving OpenAI systems. Speaking on Wednesday, he said the White House was looking at AI controls while also trying to ensure the United States remains ahead in the technology race. His remarks mark a shift in tone from an administration that has so far taken a more hands-off approach to AI.Trump made the comments after being asked about OpenAI tools being used to improperly breach the private technology of other companies. He said any move to tighten oversight would need... [Continue Reading]
360LiveNews | 30 Jul 2026 02:03 LONDON
