Vietnam's Deputy Prime Minister Le Tien Chau signed Decision 1528/QD-TTg on 11 August, committing the state to give basic AI skills to at least 10 million workers by 2030, build a corps of 50,000 applied specialists, and train 10,000 advanced specialists, some of them capable of developing core technology. Every higher-education institution must hold rules on AI use by 2030, and 80% of vocational diploma programmes must carry AI competencies matched to each trade. 1
Turkey followed a week later. A Presidential Circular of 18 August launched a 2026-2030 AI Action Plan carrying 25 billion lira, whose Worker Transformation Programme promises sector-specific training for at least 100,000 workers within 24 months and assessment of a million workers for professional transformation by 2030. 2
Both documents share a structural feature that distinguishes them from the western debate. They set headcount targets against dates, assign them to named ministries and attach them to the vocational system, which makes them auditable by anyone willing to count graduates in 2030. Neither government waited for a measurement of domestic AI exposure before committing, and neither has an instrument capable of producing one. Employer-backed upskilling in the United States moved the other way across the past year, falling from 35% of firms to 26% .
The World Bank places both countries in the group facing roughly a third of the automation risk that high-income economies carry. The states with less exposure are legislating; several of the states with more exposure are still arguing about the statistics. Vietnam's decision reads as industrial policy rather than as labour protection: a country whose comparative advantage is cheap skilled labour is trying to price that labour up before the models arrive, and 2030 is the deadline it has set itself to do it.
