A controversial AI-generated anti-drug public service announcement infused with K-pop styling has been pulled from public circulation by Hong Kong’s Correctional Services Department, following swift public criticism that the video’s creative choices inadvertently made illegal drug use look attractive to audiences.
The short educational video, designed to warn young residents about the dangers of substance abuse, leveraged artificial intelligence to replicate the high-energy, visually vibrant aesthetic that defines global K-pop music videos. Instead of resonating with its target youth demographic as intended, however, the production sparked immediate pushback from commentators and community members who flagged a critical flaw: the AI-generated visual framing of illicit substances carried a glamorous, appealing tone that ran directly counter to the campaign’s core public safety mission.
In response to the growing backlash, department officials confirmed they had taken the video offline. The incident has sparked new conversations about the unforeseen risks of using generative AI for public service messaging, particularly when creating content targeted at impressionable young audiences. While AI tools offer cost-effective and rapid content creation for government outreach campaigns, this case highlights how unvetted AI outputs can distort core messaging and produce outcomes that undermine the very goal public health and safety officials aim to achieve.