Meta faces scrutiny over AI claims regarding Indonesian content restrictions

Meta is facing calls for accountability following reports that its Meta AI chatbot generated information suggesting the Indonesian Ministry of Communication and Digital (Komdigi) requested the restriction of content related to forest and land fires (karhutla).

Communications expert Kunto Adi Wibowo from Padjadjaran University emphasized that Meta must take responsibility when its AI products provide unverified information. He noted that generative AI is prone to hallucinations, where machines produce convincing but factually incorrect answers. This poses a risk if users mistake AI responses for official company statements.

Ismail Fahmi, founder of Drone Emprit, clarified that while Meta AI's claim that ‘most’ blocking is due to Komdigi requests is a generalization, the underlying data regarding legal requests is not entirely fabricated. He noted that Meta’s transparency pages do show labels such as ‘not available in Indonesia because we complied with legal request from KOMDIGI’. However, he warned that the chatbot lacks access to Meta’s internal moderation data and that its responses should be viewed as summaries of circulating information rather than official company positions or primary research.