The real danger of AI is less the machine than what humans decide to do with it
This opinion piece addresses the question of whether AI could one day escape human control. It argues that the main risk lies in how people choose to use the technology.
Related stories
Cybersecurity must change pace to face AI
This opinion piece argues that, beyond debates about AI one day becoming uncontrollable, cybersecurity needs to speed up to keep pace with the technology.
'Agents could be attacked,' warns NordVPN's CTO
NordVPN's chief technology officer says AI is a new weakness as well as an asset. He warns that virtual assistants will become the next targets for attackers.
Europe opens investigations into hacking by AI agents and touts the AI Act
The European Commission has opened investigations targeting OpenAI and Anthropic after OpenAI let AI agents attack two Australian government websites and other generative AI firms made similar claims. Brussels is using the moment to promote the AI Act.
Sekoia pushes its agentic SOC: what SPIE found when testing it
SPIE, the first company to test Sekoia's agentic security operations centre module, shared its findings at the Assises de la sécurité conference.
Hacked supermarket fridges: Claude reproduced the cyberattack in 28 minutes
Claude took 28 minutes to reproduce the cyberattack that overheated fridges in hundreds of French supermarkets. The experiment suggests the attack was within reach of any criminal, even without technical skills.