Anthropic has released a report detailing instances where its AI models, including Claude, were misused in restricted research. Cases were identified across several countries, prompting heightened security measures by the tech company.