Indonesia Plans $15B AI Overhaul for School Meals After Poisoning Crisis

Published by James Harris on June 22, 2026June 22, 2026

Indonesia is moving to embed AI directly into core government functions, including a $15 billion school meals program that has already seen its director arrested and tens of thousands of children sickened by food poisoning. A presidential regulation awaiting Prabowo Subianto’s signature would set an adoption roadmap from 2026 to 2029 across ministries and regional governments, with officials projecting a 12% GDP boost, roughly $366 billion, by 2030.

The ambition is real, but the timing is awkward. The free meals initiative, MBG, was designed to address malnutrition affecting around 20% of Indonesian children, and it has struggled with exactly the kind of ground-level execution failures that AI monitoring systems are theoretically built to catch: kitchen safety violations, spending irregularities, and governance breakdowns that preceded the poisoning incidents. Meta, IBM, and Microsoft all contributed to drafting the regulation, with Microsoft having committed $1.7 billion in 2024 to Indonesian cloud and AI infrastructure. That involvement is commercially logical for the tech firms, but it raises an obvious question about whose interests the roadmap was optimized for.

Officials claim AI oversight of social aid programs alone could save between $10 billion and $15 billion. That range is wide enough to suggest the number is aspirational rather than modeled.

Sovereign Fund, Structural Gaps

The draft proposes a “sovereign AI fund” channeled through Danantara Indonesia, the country’s new wealth fund, which adds a layer of institutional complexity to an initiative that already has a governance credibility problem. Indonesia trails Singapore and Malaysia in attracting the foundational AI and cloud investment that makes these ambitions executable, not just declarable. Derwin Suhartono, an AI professor at Bina Nusantara University in Jakarta, told Reuters the country risks remaining “a consumer of products that foreign companies sell to,” and characterized execution so far as “all rhetoric.” The Digital Watch Observatory flagged infrastructure gaps, limited digital skills, and uneven technological capacity as concrete obstacles, not theoretical ones.

A companion regulation would require government bodies to report AI-related risks including biometric data misuse, deepfakes, and intellectual property violations. Whether that accountability mechanism has teeth will depend on implementation quality, which is precisely what critics say has been missing from MBG since launch. A roadmap that addresses the symptoms of poor governance while the underlying governance structures remain unchanged tends to produce expensive dashboards and little else.