Cybersecurity expert claims his social media tips prompted Sri Lankan police to use AI images of shooting suspect; questions remain

In an unusual, yet unconfirmed, turn of events blending public input and Artificial Intelligence (AI), Sri Lanka’s Criminal Investigation Department (CID) has reportedly released AI-generated images of a shooting suspect following a technical tip-off shared via social media.

Cybersecurity and AI policy expert Asela Waidyalankara, a Sri Lankan, claims the police action was prompted by a visual demonstration he posted on X (formerly Twitter). After seeing a hand-drawn suspect sketch shared by Sri Lanka Tweets (and claimed to have been released by ‘Police Media’) in connection with the May 16 shooting in Colombo’s Kotahena suburb, Waidyalankara used ChatGPT to produce a clearer, photorealistic image from the original.

Tagging Police Media—the Sri Lankan police’s official handle on X—he had encouraged authorities to explore how AI could support investigative efforts. ‘Perhaps it’s time for @SL_PoliceMedia to explore how generative AI can support investigative efforts’, he wrote on May 27 sharing the enhanced image.

To his surprise, the CID released AI-generated images of the wanted suspect to the public on June 2—except that he learnt of it via a report on Daily Mirror Online’s website. In a LinkedIn post reflecting on the development, Waidyalankara noted that it was ‘an encouraging sign that the message landed and that action was taken’, adding ‘it’s exactly how innovation takes root: one nudge, one use case, one adoption at a time.’

The incident has sparked fresh discussion on the role of AI in modernising policing methods. One Information Technology (IT) expert, Shiyam A., thought that it was ‘an interesting use of AI’, but he warned, ‘If the AI isn't trained on good, diverse data, it could end up creating the wrong picture altogether. That’s a big problem when we are talking about law enforcement, we don’t want innocent people getting misidentified or targeted because of a biased algorithm’.

This is a screenshot of Asela Waidyalankara’s 2 June 2025 LinkedIn post, which includes a screenshot of a ‘Sri Lanka Tweets’ post on X, his own X response, and links to both the X thread and a related Daily Mirror report.

This is a screenshot from X on 27 May 2025 showing a hand-drawn sketch of the Kotahena shooting suspect shared by ‘Sri Lanka Tweets’, followed by AI-generated images based on the sketch, created and posted in response by cybersecurity expert Asela Waidyalankara, who tagged the official Sri Lankan police handle.

Shiyam’s concern is shared many other LinkedIn commentators—most of them also from the IT community. However, there are other commentators, mostly from other sectors, who think that this is a great step that should be adopted by law enforcement, including the United States’ Federal Bureau of Investigation (FBI).

Importantly, though, the Sri Lankan police has not mentioned Waidyalankara’s tips to have prompted it to release these AI-generated images of the wanted suspect; moreover, the police’s handle on X has also not shared either the AI-generated images or the hand-drawn sketch of the suspect yet. The hand-drawn sketch was shared on X by ‘Sri Lanka Tweets’ in a post that had tagged the official handle of Sri Lankan police.

Thoughtfox has thus posed questions to Waidyalankara on his claims regarding his tips and the police’s utilization thereof to create and circulate their AI-generated images of the suspect.

Either way, using AI-generated images of wanted suspects is an interesting idea—and a dangerous one, too.