Indonesia may ban xAI’s Grok over sexualized content

Indonesia’s Ministry of Communication and Digital (Kemkomdigi) is investigating reports that Grok AI on X has been used to create and distribute sexually explicit content, including manipulated personal photos without consent.

Initial findings by Kemkomdigi indicate that Grok AI lacks adequate safeguards to prevent the production and sharing of pornographic material using real photos of Indonesian citizens.

The ministry warned that non-compliance, or lack of cooperation from tech providers, could lead to administrative sanctions or blocking access to Grok AI and X in Indonesia.

Authorities say this raises concerns over privacy violations and misuse of personal images.

Kemkomdigi is working with electronic system providers to strengthen content moderation, prevent the creation of explicit deepfakes, and establish rapid response procedures for privacy breaches.

Since January 2, 2026, Indonesia’s new criminal code enforces penalties, including imprisonment and fines, for producing or sharing pornographic or manipulated images without consent.

🔗 Source: Komdigi

🧠 Food for thought

Implications, context, and why it matters.

Indonesia’s Electronic System Provider (PSE) record makes blocking X and Grok AI feasible and politically likely

In July 2022 Indonesia blocked PayPal, Steam, Yahoo, and Epic Games for missing the PSE registration deadline 1.
A national firewall with deep packet inspection and Domain Name System manipulation funnels traffic through licensed providers, so Kemkomdigi can apply blocks across Internet Service Providers 2.
Platforms must remove unlawful content within 4 hours for urgent cases or 24 hours otherwise, with fines up to $33,000 per violation and service blocking 3.
Regulations define prohibited content broadly, including material that disturbs public order or breaks Indonesian law, which gives wide takedown latitude 2.
Public backlash followed past blocks, yet access returned after platforms registered, which confirms credible enforcement despite costs 3.

AI safety vendors should design for Indonesia’s 4-hour emergency takedown Service Level Agreements (SLAs)

Companies building deepfake detection, content watermarking, or AI safety tools can win business in Indonesia, where Kemkomdigi can order removals within 4 hours for urgent violations 3.
PSE rules require incident reporting and system integrity monitoring to Kemkomdigi, so teams need dashboards that track response times and log regulator requests 4.
Vendors with rapid moderation Application Programming Interfaces (APIs) or AI-generated content flagging can promise sub-4-hour responses that match emergency protocols.
The rules also require data and system access for Kemkomdigi on request 2, so privacy-preserving audits can offer transparency without exposing user data.
Takedown volume is high, with Google receiving over 500,000 URL delisting requests from June to December 2021 3, so scalable automation is necessary.