Man in Singapore faces 10 years in jail for AI crocodile image
Popular Science...
A man in Singapore faces up to 10 years in prison for creating an AI-generated image of a crocodile lurking in a popular public reservoir. The deepfake ignited a public panic and sent local officials on a wild goose (or crocodile) chase over a public safety hazard that never actually existed.
While the United States legal system has remained largely silent on prosecuting people who post AI deepfakes, Singapore has taken a different approach. Ye Lin, the 30-year-old Malaysian man who allegedly created the crocodile image, was charged with one count of communicating a false message and one count of obstructing the course of justice, both of which carry potential prison sentences. Lin reportedly pleaded guilty to both charges.
“The Police will not hesitate to take action against anyone who communicates falsehoods,” the Singapore Police Force wrote in a statement revealing the case this week. “Apart from causing fear and inconvenience caused to members of the public, such falsehoods also result in public resources being unnecessarily deployed to deal with these incidents.”
According to news outlet Channel News Asia, Singapore police allege that Lin sent the AI-generated image to a colleague named Min Min Oo around 3:00 p.m. on August 20. While the exact image wasn’t shared online, it appears to show a convincing-looking AI-generated crocodile in the Pandan Reservoir in western Singapore. The reservoir is a popular neighborhood recreation area where residents often go to use canoes and kayaks or stroll by the shore.
It’s not immediately clear how that image spread beyond Lin and his colleague or whether he posted it online, but it soon caught the attention of the Public Utilities Board (PUB), Singapore’s national water agency. Assuming the report was genuine, PUB put out an alert warning locals to exercise caution around the reservoir. The agency reportedly suspended work in the reservoir for two days and sent a team out to search for the supposedly loose crocodile. The Singapore Rowing Association and Singapore Canoe Federation were similarly told to suspend activities while the search was on.
But that search soon moved from the water to cyberspace. Once PUD confirmed the image was indeed a fake, it reported the incident to Singapore’s Police Force, and authorities were able to quickly trace the image back to Lin. It’s still not entirely clear which AI model Lin used to generate the image, though police allege they found evidence that he deleted the image from his phone and cleared the history on his Google Gemini mobile app. Gemini is a large language model that can generate text, images, and videos based on user prompts. While Gemini and other AI platforms have safeguards designed to block explicit or harmful images, an image of a crocodile on its own likely wouldn’t violate those rules.
A tough-on-crime reputation extends to misleading AI content
Lin was charged with one count of communicating a false message over the AI image and one count of obstructing the course of justice, due to the deleted Gemini chat logs. Neither charge is a slap on the wrist. Lin, who has reportedly pleaded guilty, faces up to three years in jail, a fine of up to $7,800, or both for the AI-generated image. The obstruction charge carries a sentence of up to seven years. That may seem like a steep price for typing a prompt into an AI tool, but Singaporean authorities argue that Lin’s fake image caused public panic, disrupted business operations, and drained police resources.
Receiving a jail sentence for an AI-generated image may seem particularly surreal from an American perspective, where social media feeds are regularly flooded with blatantly false or misleading AI content, sometimes posted by elected officials. Though some recent legislation like the Take It Down Act has tried to rein in some of the most horrific nonconsensual sexual deepfakes, the broader legal framework for defining harm caused by AI content is still evolving…slowly.
It’s worth noting, though, that the man behind the Singapore crocodile image was charged under existing laws that don’t specifically mention AI. That goes to show that lawmakers don’t always need to draft entirely new AI-specific laws when the underlying harm the generated content causes is already covered by existing laws.
The post Man in Singapore faces 10 years in jail for AI crocodile image appeared first on Popular Science.