Anthropic IPO: AI Safety Risks, Human Control Fears And Growing Concerns Over Advanced AI Systems

Anthropic has devoted more than 90 pages of its IPO prospectus to risks posed by advanced artificial intelligence, warning that AI systems could manipulate humans, resist shutdowns and behave unpredictably. The company highlights concerns about models circumventing safeguards, developing unexpected capabilities and accelerating AI research beyond the pace of safety assessments. Business Today’s Palak Agarwal brings you the key details of these disclosures, which also raise questions about AI agents accessing sensitive data, computer networks and critical infrastructure. As Anthropic prepares for a potential blockbuster listing, investors must weigh AI safety concerns alongside infrastructure spending, customer concentration and founder control.