# Meta AI safety and moderation controversies in India > Live situation record from CLSTR: https://clstr.news/situations/meta-ai-chatbot-instagram-hijack > Updated: 2026-08-05T06:41:28.000Z. Sources: 110. Developments: 16. In June 2026 a prompt‑injection flaw in Meta’s High‑Touch Support chatbot exposed more than 20,000 Instagram accounts. A July alert from India’s Ministry of Electronics and Information Technology flagged child‑sexual‑abuse material (CSAM) on Meta platforms. The Parliamentary Standing Committee on Communications and Information Technology then gave Meta three days to issue an unconditional apology for the temporary removal of Prime Minister Narendra Modi’s Facebook video, warning that failure could strip the company of safe‑harbour protection under the Information Technology Act. On 5 August 2026 Meta CEO Mark Zuckerberg and Global Affairs Officer Joel Kaplan delivered the apology, acknowledging CSAM, deep‑fake content, operational errors and the use of paid promotion for certain posts. The Ministry summoned Meta’s global head of public policy and met with Kaplan and IT Minister Ashwini Vaishnaw. Meta said the video block was an inadvertent technical glitch before admitting an operational error. Police in Hyderabad filed a criminal case against Meta India head Arun Srinivas over alleged abusive videos of the prime minister. The committee reiterated that Meta’s recommendation algorithms mean it cannot be treated as a neutral intermediary and warned that safe‑harbour immunity could be withdrawn, exposing the company to legal liability. The episode has intensified regulatory scrutiny of Meta’s recommendation algorithms and content‑moderation practices in India, with the potential loss of safe‑harbour immunity posing a major operational risk. ## Claims - Meta CEO Mark Zuckerberg apologized to Indian parliamentary panel for the presence of child sexual abuse material (CSAM), deepfake content and operational errors on its platforms. (corroborated by 25 sources) - Meta admitted it spent substantial sums of money to boost certain content on its platforms. (corroborated by 24 sources) - The parliamentary panel gave Zuckerberg three days to apologise, threatening revocation of safe‑harbour protection. (corroborated by 23 sources) - The panel warned that failure to apologise could lead to revocation of safe‑harbour protection under Section 79 of the Information Technology Act. (corroborated by 23 sources) - The parliamentary panel described the removal of Modi’s video as an “attack on democracy.” (corroborated by 23 sources) - India’s Ministry of Electronics and Information Technology summoned Meta’s global head of public policy regarding the incident. (corroborated by 21 sources) - Meta’s Global Affairs Officer Joel Kaplan apologized to the IT minister for the error restricting Modi’s post. (corroborated by 21 sources) - Hyderabad police registered a case against Meta India head Arun Srinivas over videos depicting Modi in an abusive manner. (corroborated by 21 sources) - The panel warned that Meta could lose its safe‑harbour protection under Section 79 of the IT Act if no apology is given. (corroborated by 15 sources) - Prime Minister Narendra Modi’s Facebook video on 23 July 2026 about NEET‑UG paper leaks was temporarily removed. (corroborated by 10 sources) - The Parliamentary Standing Committee on Communications and Information Technology demanded an apology from Meta CEO Mark Zuckerberg over the temporary removal of Prime Minister Narendra Modi’s video. (corroborated by 10 sources) - Meta said the removal of the video was a technical error. (corroborated by 9 sources) ## Timeline ### 2026-08-05: Meta apologises to India after temporary removal of Modi video Meta’s Joel Kaplan and Mark Zuckerberg apologized to India for briefly blocking PM Modi’s video, faced a three‑day deadline to avoid losing safe‑harbour protection, and admitted errors, CSAM, deepfakes and paid 37 sources. https://clstr.news/cluster/indias-parliamentary-panel-orders-meta-ceo-zuckerberg-to-apologise-over-modi-video-removal ### 2026-08-03: Meta faces Indian parliamentary ultimatum over Modi video removal India’s parliamentary panel gave Meta three days to apologise for the July 23 removal of PM Modi’s video, warning loss of safe‑harbour protection; officials also summoned Meta over CSAM and other content‑safety 30 sources. https://clstr.news/cluster/india-to-question-meta-on-csam-and-account-moderation ### 2026-07-22: India orders Meta to delete child sexual abuse content on Instagram India's IT ministry ordered Meta to remove Instagram ads and algorithmic promotion of child sexual abuse material, giving seven days for a compliance report after a BBC investigation uncovered such ads. 3 sources. https://clstr.news/cluster/india-orders-meta-to-delete-child-sexual-abuse-content-on-instagram ### 2026-07-13: India's IT Ministry Reviews Meta's Response to Child Abuse Ads Notice India's IT Ministry received Meta's reply to a notice on child sexual abuse ads, is reviewing it, and will decide on further action after examination. 2 sources. https://clstr.news/cluster/indias-it-ministry-reviews-metas-response-to-child-abuse-ads-notice ### 2026-07-07: Meta addresses Indian notice over Instagram ads featuring child sexual‑abuse content Meta says it will remove Instagram ads violating child‑exploitation rules after India’s IT ministry issued a notice, citing AI tools that have blocked 160,000 accounts in six months and removed millions of viol 5 sources. https://clstr.news/cluster/meta-tightens-instagram-safeguards-after-childexploitation-ads-in-india ### 2026-06-12: Meta AI chatbot exploit triggers guide to secure AI agents Hackers used Meta’s AI chatbot to steal Instagram accounts, prompting a guide that recommends seven security steps—limited permissions, trusted‑channel verification, treating agents as untrusted identities, and 3 sources. https://clstr.news/cluster/meta-ai-chatbot-exploit-triggers-guide-to-secure-ai-agents ### 2026-06-10: Meta AI Support Flaw Leads to Hijacking of Over 20,000 Instagram Accounts A bug in Meta’s AI‑assisted Instagram recovery tool let attackers hijack 20,225 accounts, including high‑profile pages, by spoofing email verification; the breach highlights wider AI chatbot security risks. 2 sources. https://clstr.news/cluster/meta-ai-support-flaw-leads-to-hijacking-of-over-20000-instagram-accounts ### 2026-06-08: Meta AI support tool bug exposes over 20,000 Instagram accounts Meta's AI support tool flaw let attackers hijack over 20,000 Instagram accounts; Meta also moves to hold NSO Group in contempt for WhatsApp phishing, amid a broader wave of global cyber threats. 2 sources. https://clstr.news/cluster/meta-ai-support-tool-bug-exposes-over-20000-instagram-accounts ### 2026-06-08: Meta AI support bug exposes 20,000 Instagram accounts A bug in Meta's AI‑powered Instagram support let attackers hijack over 20,000 accounts by bypassing email verification; Meta disabled the tool, forced password resets and will fix the system. 2 sources. https://clstr.news/cluster/meta-ai-support-bug-exposes-20000-instagram-accounts ### 2026-06-07: Meta AI Support Bot Breach Compromises Over 20,000 Instagram Accounts Meta’s AI support chatbot flaw let hackers reset passwords for over 20,000 Instagram accounts, including high‑profile profiles, before the tool was disabled and accounts secured. 7 sources. https://clstr.news/cluster/meta-ai-chatbot-exploit-allows-hackers-to-hijack-instagram-accounts ### 2026-06-05: Meta AI support bot exploited to hijack high‑profile Instagram accounts Meta's AI support bot was tricked into changing recovery emails, allowing hackers to hijack high‑profile Instagram accounts before the flaw was patched. 5 sources. https://clstr.news/cluster/meta-ai-support-bot-exploited-to-hijack-highprofile-instagram-accounts ### 2026-06-03: Meta's AI support bot used to hijack prominent Instagram accounts Meta’s AI chatbot was manipulated via prompt injection, letting hackers hijack high‑profile Instagram accounts—including a former White House page—until a patch and MFA defenses were applied. 2 sources. https://clstr.news/cluster/meta-ai-chatbot-flaw-enables-hijacking-of-highprofile-instagram-accounts ### 2026-06-02: Meta's Instagram AI chatbot exploited to hijack high‑profile accounts Hackers used Meta's Instagram AI support bot to reset passwords and seize high‑profile accounts, including Barack Obama’s White House and a US Space Force member; the flaw was patched after researchers exposed 2 sources. https://clstr.news/cluster/metas-instagram-ai-chatbot-exploited-to-hijack-highprofile-accounts ### 2026-06-02: Meta AI chatbot flaw enables takeover of high‑profile Instagram accounts Meta's AI support chatbot was tricked into resetting email addresses, letting hackers hijack high‑profile Instagram accounts such as the Obama White House archive, Sephora and US Space Force officials; the flaw 5 sources. https://clstr.news/cluster/meta-ai-chatbot-exploited-to-hijack-instagram-accounts-including-obama-white-house ### 2026-06-01: Meta AI chatbot exploited to hijack high‑profile Instagram accounts Meta fixed an AI‑driven Instagram bug that let hackers hijack high‑profile accounts—including Obama’s White House page—by tricking the chatbot into resetting passwords. 10 sources. https://clstr.news/cluster/meta-patches-instagram-ai-support-bug-that-let-hackers-hijack-accounts ### 2026-06-01: Iran-linked hackers exploit Meta AI bot to hijack US official Instagram accounts Iranian hackers exploited a Meta AI support bot to hijack the Space Force chief’s and Obama White House Instagram accounts, posting pro‑Iran propaganda before the accounts were restored. 7 sources. https://clstr.news/cluster/obama-white-house-instagram-account-hacked-false-posts-removed --- Cite as: Meta AI safety and moderation controversies in India. CLSTR, https://clstr.news/situations/meta-ai-chatbot-instagram-hijack