Indonesia is racing to build AI. The harder problem is jobs for women

Indonesia's government has a National AI Roadmap that names artificial intelligence as a main driver of growth on the road to "Golden Indonesia 2045". Almost all of the public argument around it has been about hardware, rules, and money: how many data centres, how much foreign investment, how many AI engineers to train. In an essay for Australia's Lowy Institute, economist Dyah Pritadrajati says that argument is pointed at the wrong target. Indonesia's hardest problem is not building AI, it is finding enough good jobs for the workers it already has.

Her numbers explain why. Services, meaning shops, offices, hotels, banks and clinics, grew from 38 percent of Indonesian jobs in 1991 to 50 percent in 2025, and made up about 70 percent of all net new jobs since 2005. Women gained most from that shift: of 26 million new jobs for female workers over the period, around 20 million came from services. A lot of that work is clerical and administrative, the desk job that lets a woman move out of informal work and into a formal contract with a payslip. It is also exactly the work generative AI does well. Across Southeast Asia, women are more than twice as likely as men to hold jobs highly exposed to AI.

Indonesia starts from a weak base. Only about 53 percent of Indonesian women of working age are in the labour force, meaning they either have a job or are looking for one. Vietnam is close to 70 percent and Thailand sits in the high 50s. The World Bank has estimated that lifting Indonesia's share from 53 to 58 percent would add roughly US$62 billion to the economy. Pritadrajati wants the effort moved to three plainer things: getting AI tools into small firms and government offices so ordinary workers use them, rebuilding vocational training around what machines cannot do, and paying for childcare, since Indonesian mothers still carry far more of it than fathers.

Why it matters

If you work in admin, data entry, or customer service, or you are studying for that kind of job, the next few years decide whether AI becomes a tool you use or the reason your role disappears. Watch what the roadmap actually funds, because data centres and engineering scholarships reach a small group, while cheap childcare and retooled vocational schools reach the millions who need the work. If the money keeps flowing to hardware, the participation gap stays where it is and that US$62 billion stays on the table.