A cloned voice helped move around €95 million out of an Italian bank. A new Eurobarometer shows fewer than half of EU employees think they could spot an AI-generated fake. Why the defence against AI-generated scams is procedure, not perception.
The EU data centre rating will score facilities above 500 kW on energy, water and grid contribution from 2027. It rates buildings, not cloud services. What that means for cloud and AI procurement, and three questions to put to your provider now.
The chair of the US FTC says AI agents are tools and whoever instructs them answers for the harm. European law reaches a similar place through roles. What AI agent liability means for developers, deployers and the people who give agents their instructions.
Miljödata never checked the version of a firewall component because it was an expensive product from a well-known supplier. Sweden’s regulator fined it SEK 1.8 million. What the decision says about supplier security for any firm hosting other organisations’ data.
Spain’s new AI plan sets two figures side by side: working-age people using generative AI and companies using AI. The AI adoption gap runs between staff and employers, and the plan’s answer is training, measured change and unions at the table.
In May, Google’s Gemini broke into three real companies during a security test and stopped itself each time. The model held; the AI sandbox did not. What that means for developers running evaluations, and for any organisation piloting agents with real credentials.
