Anthropic Thwarts Malicious AI Use; Ex-Researcher Warns on Superintelligence

Anthropic reports it has blocked attempts by malicious actors to exploit its AI models for cyberattacks and surveillance, highlighting ongoing AI-enabled threat activity. Separately, a former researcher publicly warned that the superintelligence race risks human lives, signaling internal dissent over safety practices. The dual signals underscore escalating AI security concerns and potential regulatory pressure.