Sri Lanka Police have issued a public warning about artificial-intelligence deepfake scams, saying criminals are using the technology to impersonate people the victim trusts in order to extract money and personal data.
The warning, reported by the Daily Mirror at 3.43 p.m. on Wednesday, cites INTERPOL’s assessment that AI is being used increasingly in fraud schemes to manufacture convincing fake video and voice recordings of real individuals.
The mechanism the police describe
The specific claim in the warning is about how little raw material an attacker needs.
Police say photographs or short voice recordings already posted on social media are sufficient to generate deepfake content imitating family members, friends and other trusted contacts. Those fabricated identities are then used to push victims into transferring money, disclosing banking details or handing over confidential information.
That is a meaningful detail, because it shifts the exposure. The risk being described does not depend on a person being wealthy, prominent or targeted in advance — it depends only on their voice or face being available online, which for most social media users it is.
What the police advise
The guidance is procedural rather than technical, and it does not ask the public to detect fakes by eye or ear:
- Treat unexpected calls, messages and video calls from relatives, friends or trusted contacts with caution, particularly where there is a sudden request for money or sensitive information.
- Verify identity through a separate, trusted channel before acting — calling a known number back, rather than replying through the channel the request arrived on.
- Never share banking credentials, passwords or verification codes in response to a suspicious approach.
- Be cautious with unfamiliar digital content generally, since AI-generated audio and video is becoming more realistic.
The verification step is the substance of the advice. It works regardless of how good the fake is, which the “look for visual glitches” guidance common in earlier public warnings does not.
Context
The warning follows two strands this desk has covered. In August, Deputy Minister Abeysinghe warned that scam callers were using his name to approach members of the public. In September, the Prime Minister said deepfake images were being used to target women political leaders.
Wednesday’s advisory is broader than either: it concerns ordinary members of the public rather than public figures, and financial fraud rather than reputational attack.
Not reported
The Daily Mirror’s filing does not name the police division that issued the warning or any officer, and gives no reporting channel or hotline for victims — a notable omission in an advisory that asks the public to recognise a crime in progress.
It also gives no Sri Lankan case data: no number of reported deepfake fraud incidents, no losses, no arrests, and no indication whether the warning responds to cases already recorded here or anticipates a threat observed elsewhere. The INTERPOL material is cited as the basis, which leaves open whether any local incidents have yet been investigated.