Current Affairs · · GS3 · Internal Security

AI-generated fake videos fuel Ujjain clashes, testing platform accountability

Police say AI-generated videos falsely showing the whole Shahi Masjid in Ujjain being demolished helped mobilise a crowd that clashed with police on 28 September 2026. Only a road-affected portion was being removed. The case tests India’s rules on synthetic media, incitement and platform duties.

Event date:

REq1
  1. Context1 / 4
    • Clashes broke out in Ujjain, Madhya Pradesh, on 28 September 2026. A crowd threw stones at police near the Shahi Masjid. Police used tear gas and mild force.
    • The mosque stands partly on a road being widened before the 2028 Simhastha Kumbh. On 9 September the Madhya Pradesh High Court dismissed petitions against the demolition notices.
    • On the day of the clashes, the mosque side reached a settlement with the Ujjain Municipal Corporation. The mosque committee agreed to remove the road-affected portion itself.
    • Police say videos made with artificial intelligence (AI) showed the entire mosque being flattened by an earthmover. Ujjain Superintendent of Police Pradeep Sharma called that depiction completely false.
    • Police have registered eight first information reports (FIRs) against 20 people across three police stations. Five of the accused are content creators. Investigators are also examining more than 100 Instagram accounts.
  2. Key highlights2 / 4
    • Cheap, fast fakes. Generative AI cuts the cost and the time of making a convincing video during a tense moment. A fake can travel faster than an official correction.
    • Offline grievance, online amplification. The videos landed on a dispute that already existed over a legally cleared demolition. They did not create the dispute. Police say a section of the mosque community was misled by them.
    • Liability needs intent. Deliberately faking footage to provoke a crowd is different from sharing a clip in good faith, reporting on it, or criticising the demolition lawfully.
    • Platform role. Platforms hold the data on how content spreads. They can label it, slow its reach and keep evidence. But takedowns that are too broad can suppress lawful speech.
  3. Key concepts3 / 4

    Synthetically generated information (SGI). The Information Technology (IT) Amendment Rules, 2026, in force from 20 February 2026, define SGI in Rule 2(1)(wa). It covers audio, visual or audio-visual content that is made or altered using a computer resource so that it appears real, and that shows a person or an event as indistinguishable from reality.

    The test looks at the content, not the tool. A text-only fake rumour is not SGI, but it can still be unlawful information.

    Safe harbour (Section 79, IT Act, 2000). An intermediary, such as a social-media platform, is not liable for what its users post if it follows due diligence. Think of a courier company: it does not answer for what is inside a parcel, unless it ignores a lawful warning.

    The 2026 rules add duties. Platforms must label SGI, and on a valid court or government order they must remove unlawful content within 3 hours, down from 36 hours.

    Incitement and speech offences. The Bharatiya Nyaya Sanhita (BNS) punishes promoting enmity between groups (Section 196), deliberate and malicious acts that outrage religious feelings (Section 299) and statements that cause public mischief (Section 353). Courts read these provisions narrowly, and intent is central.

    Limits on speech. Any restriction must fit a ground listed in Article 19(2) of the Constitution, including public order. Orders that are vague or too broad can fail this test.

  4. Way forward4 / 4
    • Rapid multilingual fact-check cells. Police and district administrations need a fast, official channel to say what is true, in local languages, within hours.
    • Provenance labels. Require visible labels and embedded metadata, the hidden data that travels with a file, on synthetic political and communal media. Then test how well platforms follow the rules.
    • Keep audit trails. Platforms should retain records of takedowns and of how content spread, so investigators can trace where a clip began without guesswork.
    • Narrow, reasoned orders. Takedowns should target specific content and give written reasons. Blanket blocking risks silencing lawful criticism.
    • Community liaison and media literacy. Peace committees and local leaders can calm a crowd before a rumour hardens. Public campaigns can teach people to pause before they share.

Sources

Syllabus

PaperSubjectSub-topic
GS3Internal SecurityChallenges to internal security through communication networks; basics of cyber security
GS2GovernanceGovernment policies and interventions; statutory and regulatory bodies
GS1SocietyCommunalism as a challenge to social harmony

Topics

India Contemporary EventsFundamental RightsGovernanceInformation and Communication Technology

Practice questions

  1. Consider the following statements about the Information Technology (Intermediary Guidelines and Digital Media Ethics Code) Amendment Rules, 2026: 1. They define synthetically generated information by the realistic appearance of the content, not only by the use of AI tools. 2. They lengthened the time for removing unlawful content on a government order from 3 hours to 36 hours. Which of the statements given above is/are correct?

    1. 1 only
    2. 2 only
    3. Both 1 and 2
    4. Neither 1 nor 2
    Show answer

    Answer: A. Statement 1 is correct: the definition rests on content that appears real and is indistinguishable from real persons or events, whatever tool made it. Statement 2 is wrong, because the window was shortened from 36 hours to 3 hours, not lengthened.

    Difficulty: medium · statement

  2. With reference to the safe harbour protection available to intermediaries in India, consider the following statements: 1. It arises under Section 79 of the Information Technology Act, 2000. 2. It is available irrespective of whether the intermediary observes due-diligence requirements. 3. It protects an intermediary from liability for third-party content. How many of the statements given above are correct?

    1. Only one
    2. Only two
    3. All three
    4. None
    Show answer

    Answer: B. Statements 1 and 3 are correct: Section 79 of the Information Technology Act, 2000 shields an intermediary from liability for content posted by others. Statement 2 is wrong, because the protection depends on observing due diligence, and the 2026 amendment rules add labelling and a 3-hour takedown duty to that due diligence.

    Difficulty: medium · statement

  3. Which one of the following is a ground for restricting free speech under Article 19(2) of the Constitution of India?

    1. Public order
    2. Political expediency
    3. Administrative convenience
    4. Economic stability
    Show answer

    Answer: A. Article 19(2) lists public order among the grounds on which reasonable restrictions may be placed on freedom of speech and expression. Political expediency, administrative convenience and economic stability are not grounds listed in it.

    Difficulty: easy · statement

Mains practice

Answer-writing practice on this article. Attempt it first, then open the hints.

  1. GS3 · 250 words

    "Verification has not kept pace with generation." Examine how AI-generated content threatens internal security, and evaluate India’s regulatory response to it.

    Show hints
    1. Define synthetic media and explain why generative tools make convincing fakes cheap and fast.
    2. Show how such content spreads faster than an official correction and can turn an existing local grievance into a crowd.
    3. Explain the 2026 amendment rules: the definition of synthetically generated information, labelling duties and the 3-hour takedown window.
    4. Discuss safe harbour under Section 79 and how conditioning it on due diligence shifts the burden to platforms.
    5. Weigh the free-speech risk of broad takedowns, then argue for rapid official fact-checking, provenance labels and narrow reasoned orders.
  2. GS2 · 150 words

    Critically examine the balance between platform accountability and freedom of speech in India’s approach to deepfakes.

    Show hints
    1. Set out the limits in Article 19(2), especially public order, and the rule that restrictions must be narrow and reasoned.
    2. Argue the case for faster takedowns against the cost to due process when a platform removes content in 3 hours.
    3. Explain how labelling and provenance metadata can reduce harm without deleting lawful speech.
    4. Conclude on safeguards: written reasons, specific targeting and preserved audit trails for review.