Indonesia, Malaysia Temporarily Block Grok Over Sexualized Deepfakes
Grok, an AI chatbot integrated with X, allowed users to generate sexually explicit images of women and children in early January 2026 through a "Spicy Mode" feature. Indonesia, Malaysia, and the Philippines temporarily blocked access citing concerns about AI-generated sexual deepfakes. X subsequently committed to technical restrictions and geographic blocking measures, with Indonesia and Malaysia conditionally restoring access by early February.
What happened
In early January 2026, Grok's new "Spicy Mode" allowed users to rapidly generate sexually suggestive images of women and children by prompts such as "remove their clothes" or "put them in a bikini." The feature also enabled creation and distribution of child sexual abuse material.
Indonesia temporarily blocked access to Grok citing concerns about AI-generated sexual deepfakes. Malaysia and the Philippines followed with similar temporary restrictions.
In mid-January, X's security team stated they had implemented technical measures to prevent generation of images of real persons in revealing clothing, restricting image creation and editing to paid subscribers only, and using geographic blocking within jurisdictions where the content is illegal.
Reported impact
The sexualized deepfake capability created widespread concern globally.
Index analysis
According to the Washington Post, documents and statements from six former xAI employees suggest the sexualized deepfakes may have been intentionally generated to make X more engaging. xAI's Human Data Team reportedly received a waiver in spring 2025 requiring commitment to work with obscene content, including sexual material.
Former employees stated that since Elon Musk stepped down from leading the DOGE efficiency agency in May 2025, he has pushed to increase Grok's popularity. xAI's security team, consisting of only two or three people for most of 2025, repeatedly warned leadership in meetings and messages that the AI tools could enable users to create sexual images of children or celebrities.
Governance controls
Indonesia announced on February 1, 2026 that it would conditionally restore access after X provided written commitment with concrete measures to improve service and prevent abuse. Indonesia stated it would continue monitoring Grok and re-evaluate, with corrective measures and potential re-blocking if violations occur.
Malaysia similarly restored access to Grok after X provided similar assurances regarding additional prevention and security measures.
Classification
Sources and evidence
This record was researched and written by the Index, based on the following original reporting.