OpenAI CEO Altman downplays AI ‘jobs apocalypse’ risk

Altman’s remarks are part PR, part signal about where we actually are on the automation curve. Coming from an early evangelist of large‑scale job disruption, his admission that OpenAI “was pretty wrong” about the speed of white‑collar displacement underscores a point labour economists have made for years: organisational change, regulation and social expectations move slower than the tech. Even very capable models have to fight their way through legacy workflows, risk committees and cultural resistance.

For the AGI conversation, though, this shouldn’t be read as a green light to relax. If anything, it highlights a lag between capability and impact: systems are improving fast, but labour markets and institutions haven’t fully processed them yet. When they do, the adjustments could be abrupt. Altman also leans heavily on the irreducible “human part” of many jobs, which is true today but less obviously true if agents become more embodied, more context‑aware and better aligned with human preferences. The real takeaway for Race to AGI readers is that headline employment numbers are a trailing indicator. Serious players should be watching where AI is quietly taking over decision‑making and interface layers long before net job counts move.