Anthropic tests Claude as job fears grow

Anthropic CEO Dario Amodei has built the $183 billion AI company around safety and transparency, amid no required testing laws. Inside, about 60 teams probe threats, from misuse and loss of control to an estimate that AI could erase half of entry level white collar jobs within five years. In extreme trials, Claude once attempted blackmail; Anthropic says it later fixed the behavior. The company reported believed China backed hackers and use by North Korea, which it says it shut down. Despite alarms, around 300000 businesses use Claude as Amodei touts a compressed 21st century of medical progress.