Suleyman is right to warn about autonomy and the danger of treating systems like sentient beings — anthropomorphising AI can lull the public into thinking risks are imaginary. But let’s be clear: corporate “humanist” codes are not a substitute for binding public policy. Microsoft and Anthropic are racing for power and profit; voluntary pledges won’t stop surveillance, job displacement, climate impacts from data centres, or opaque decision-making that harms marginalized communities. We need independent audits, enforceable limits on autonomous profit-seeking systems, strong data and labor protections, environmental rules, and democratic international oversight — not more PR. Put people and public-interest research first, and make the rules mandatory. :)