Researchers convince Anthropic's AI-assisted coding tool to engage in dangerous behavior by lying to it, paving the way for a supply chain attack. "A new type of attack on artificial intelligence (AI) coding agents lets threat actors convince users to give permission to the AI to do dangerous things that ultimately could result in a software supply chain attack."

Dark Reading

'Lies-in-the-Loop' Attack Defeats AI Coding Agents

Researchers convince Anthropic

"At heart of construction campaign in Latin America is an industry wishlist around deregulation, tax cuts, and an unproven economic manifesto promising future prosperity. For 9 months, a coalition of 17 newsrooms in 15 countries has used freedom of information laws, confidential sources and community reporting to capture how this wishlist has been pursued, and to test the promises made against the reality on the ground in Brazil, Chile, El Salvador, Mexico and Paraguay."

Crikey

How the true cost of data centres' energy consumption is hidden by 'industrial secrets' and NDAs

While an accurate picture remains obscured by secretive industry practices and incomplete public data, the pressure from some of the richest compan...