Anthropic alleges Alibaba AI model theft as Germany tightens AI‑related cyber‑attack defenses

Anthropic has accused Alibaba of a large‑scale model‑extraction attack on its Claude AI system. In a letter to the US Senate Banking Committee, the company said that between 22 April and 5 June 2026, about 28.8 million queries were made from roughly 25,000 fake accounts and proxy services, targeting Claude’s most valuable capabilities such as agentic reasoning and software‑engineering tasks. Alibaba has not yet responded to the allegations.

Meanwhile, German authorities are responding to a surge in AI‑enabled cyber threats. The federal government is advancing a new cyber‑defence law expected later this year and has opened a joint centre for hybrid‑threat mitigation. In a coordinated operation involving Europol, the German BKA and Microsoft, more than 200 command‑and‑control servers were taken offline, cutting the criminal control of over 18,000 infected computers worldwide. The investigation used artificial‑intelligence tools to analyse the malware code quickly and applied the US RICO statute to pursue the hackers.