Could 37% of Jobs Be Replaced by AI? Andrew Ng Warns Against "Toxic Positivity"

The claim and the caveat

It has been reported that Andrew Ng, the Chinese‑American AI entrepreneur and founder of deeplearning.ai and co‑founder of Coursera, warned that as much as 37% of jobs could be replaced by artificial intelligence. The figure comes from an interview reported by ifeng (凤凰网), and Ng used it to make a broader point: enthusiasm about AI’s opportunities can slide into "toxic positivity" that downplays real displacement risks. Short sentence: optimism is not a substitute for policy.

What Ng urged — and why it matters

Ng reportedly argued that complacency will let employers and platforms exploit workers who lack the skills or protections to navigate rapid automation. He urged a balanced approach: accelerate training and deployment of useful AI, while recognising transitional costs and supporting displaced workers. This is not just a western problem. In China, major tech players such as Baidu (百度), Alibaba (阿里巴巴) and Tencent (腾讯) are racing to commercialise generative AI, raising the same questions about which jobs will be augmented and which will disappear.

Geopolitics and the labour market

Why mention geopolitics? Because AI development sits inside a wider tech rivalry between the United States and China, and trade and export controls on chips and software shape who builds what, and how fast. Policy choices — subsidies for retraining, unemployment backstops, and rules on AI deployment — will determine whether automation becomes broad-based prosperity or concentrated disruption. Who pays for retraining? Who sets guardrails? Those are political questions as much as technical ones.

Next steps

Ng’s warning is a call for clearer thinking, not panic. Governments, companies and educators must plan for transitions now. Reportedly, that means scaling reskilling programs, updating labour rules, and demanding transparency from AI deployers. The key question remains: can societies move faster than the machines they build?