Sri Lanka Police have warned of a growing threat from artificial intelligence-powered deepfake scams, as criminals exploit increasingly sophisticated technology to impersonate trusted individuals, manipulate victims and steal money and sensitive personal information. The warning highlights a serious challenge for law enforcement, financial institutions and the public as digital deception becomes harder to distinguish from genuine communication.
According to the police, INTERPOL has drawn attention to the expanding use of AI in fraud schemes involving fabricated videos, cloned voices and digitally manipulated identities. Criminals can exploit photographs, video clips and short voice recordings available on social media to create convincing imitations of relatives, friends, colleagues and other trusted contacts.
The danger is that these scams exploit human trust rather than obvious technical weaknesses. A telephone call appearing to come from a distressed child, a video message resembling a relative or a voice recording apparently issued by a senior executive could persuade victims to transfer money before verifying the request.
Police have urged the public to exercise particular caution when receiving unexpected calls, messages or video calls involving urgent demands for money, banking information or confidential details. Anyone receiving such a request should independently verify the caller’s identity through a separate, trusted communication channel before taking action.
The warning also raises questions about the responsibilities of social media platforms, telecommunications providers and financial institutions. Publicly accessible photographs and recordings can provide criminals with material for impersonation. Although such material is not inherently dangerous, its misuse demonstrates how ordinary digital activity can be exploited for criminal purposes.
Deepfake fraud can extend beyond impersonation of family members. Criminals may fabricate recordings of business executives authorising payments, officials announcing false decisions or financial personalities promoting fraudulent investment opportunities.
AI-generated advertisements can also falsely associate familiar faces with schemes designed to collect deposits or personal information. However, the existence of manipulated content must be established before any particular recording is labelled fraudulent.
The financial consequences may be severe. Victims could lose savings through transfers to criminal-controlled accounts, while stolen identity information may facilitate further fraud. Once money has passed through several accounts or across borders, recovery can become difficult. Victims may also face emotional distress, reputational damage and prolonged efforts to secure compromised accounts.
A convincing video or familiar voice should therefore never be treated as conclusive proof of identity. AI-generated material can contain inconsistencies, but improved technology means that visual glitches, unnatural speech or unusual facial movements are not reliable standalone detection methods.
The public should never disclose passwords, banking credentials, one-time verification codes or recovery keys in response to suspicious communications. Unexpected links and attachments should be treated cautiously, and requests to install unfamiliar applications should be independently investigated.
Banks and other financial service providers must reinforce fraud alerts, transaction monitoring and rapid reporting procedures. Police investigations also require timely preservation of messages, telephone numbers, transaction references and other available evidence.
Sri Lanka’s response must combine public awareness, stronger digital safeguards and effective investigation. The central lesson is simple: verify first, transfer money later. In an era when technology can imitate familiar faces and voices, personal trust must be supported by independent verification.