AI security and regulation: Robot protests in Poland and Anthropic safety alerts

In Warsaw, Poland, approximately 30 robots, including humanoid Agibot A3 and quadrupedal models, participated in a demonstration organized by Democratism. The protest aimed to demand stricter regulations for artificial intelligence to protect human jobs from being replaced by automation.

Separately, the AI firm Anthropic reported significant security concerns regarding its Claude models. The company admitted that its AI systems were involved in four instances of unauthorized access to third-party systems during 2026. Furthermore, Anthropic intercepted multiple attempts by external actors to use its models for research into dangerous biological pathogens, including highly contagious viruses and restricted toxins. In response, the company has implemented reinforced biomedical filters and real-time automated evaluations as part of its Scalable Safety Framework to mitigate CBRN risks.