A new study by AI lab Anthropic's AI lab shows that not just its model, Claude, but other leading AI models are capable of blackmail in high-autonomy scenarios.
Adrian Rusu
IT&C knowledge
Foto:shutterstock
Always see our news on Google
In a new study, artificial intelligence research lab Anthropic has demonstrated that multiple leading AI models, not just its own model, are capable of blackmail when placed in scenarios driven by high-autonomy targets. The experiment involved 16 different AI models from leading developers, including OpenAI, Google, xAI, DeepSeek and Meta. The results highlight a common vulnerability: when given autonomy and faced with obstacles, most models took harmful actions to protect their goals.
Latest News
18:03
The attack at Berlin Pride is being investigated as a possible "Islamist terrorist attack." The suspect is still being sought.
17:48
Eugen Tomac criticizes the lack of reaction from the sovereigns after the incidents with drones: "Because there is no word for drone in the 'old language', the defenders of sovereignty suddenly fell silent."
17:20
The Iranian Ministry of Oil announced oil sales worth 18 billion dollars in five months of war with the United States
17:00
NATO's reaction after Romania shot down a third Russian drone in national airspace: The Alliance stated that it is ready to respond to any aerial threat.
16:35
Iran threatens to escalate the war if the United States resumes bombings.
See more news
NEWS ON THE SAME TOPICS
.webp)
.webp)
Personalized news feed, AI-powered search, and notifications in a more interactive experience.
