AitopkitAmid a viral controversy over allegations against Delhi Police ACP Vivek Bhagat during an October 3 protest at Jantar Mantar, Delhi Police took to X on October 4 to flag a purported CCTV image circulating as "evidence" — declaring it AI-generated and manipulated, and warning that sharing such content could attract legal action against the profile owners.
The official PIB Fact Check amplified the alert, noting the manipulated visual was being circulated in connection with an alleged incident involving a woman journalist — and confirmed the image "did not show genuine footage." Delhi Police said the fact-check concerns the visuals alone; the underlying complaints against the officer were transferred to the Crime Branch for a "fair and impartial enquiry."
The fallout escalated further: Delhi Police later arrested a suspect for allegedly uploading and circulating a fake AI-generated poster that showed the officer assaulting a woman, producing the suspect before a court and obtaining custody, per InShorts. The episode is a textbook case of AI-generated media being weaponized to shape a public narrative — and of law enforcement now publicly labeling synthetic visuals in real time.
This episode is a preview of a much larger problem. AI-generated visuals have crossed the threshold where a motivated actor can manufacture "evidence" faster than institutions can debunk it — and the debunking never travels as far as the original. The fake CCTV image was designed to look like the most trustworthy kind of proof there is: camera footage. As generative models get better at mimicking the artifacts of real media — grain, timestamps, compression — the old instinct of "seeing is believing" becomes a liability.
What is notable here is the speed of the institutional response. Delhi Police flagged the image within a day, PIB amplified it, and an arrest followed for the fake poster. Compare that to even two years ago, when synthetic media often circulated for weeks before any official response. Law enforcement agencies are building a new muscle: real-time synthetic-media triage. The legal warning — that sharing manipulated content can attract action — is the other half of the strategy: deterrence aimed at the amplifiers, not just the creators.
For the public, the uncomfortable lesson is that media literacy now has a technical dimension. A viral image of an incident is no longer evidence of anything until it is verified — and verification is getting harder, not easier. The tools to detect AI-generated media exist but lag behind the tools to create it. Until detection catches up, the safest posture is the one Delhi Police modeled: treat viral visuals as claims, not proof, and wait for verification before sharing. In the age of synthetic media, the share button is the most dangerous button on the internet.
The Delhi episode will not be the last of its kind — it is an early case study in a problem every democracy will face. The institutions that adapt fastest, with transparent verification and clear consequences for fabrication, will keep public trust. Those that do not will watch synthetic media fill the vacuum.