LONDON | Hong Kong’s Correctional Services Department removed an AI-generated anti-drug video after criticism that the K-pop-style campaign risked making illegal substances look appealing rather than discouraging them.
BBC News reported the core development, and the Times of India also reported that the prison service withdrew the AI-generated video after social-media backlash. The case is a technology story because the controversy is not only about the message. It is about how public agencies use generative media, pop-culture formats and synthetic performers when the subject is sensitive and the audience may include young people.
CGN News is not reproducing the video or disputed wording. The public issue is whether AI-generated public-service campaigns are being reviewed with enough care before release.
What is known
BBC News reported that Hong Kong’s Correctional Services Department removed the video after criticism. The campaign reportedly used AI-generated K-pop-style imagery to deliver an anti-drug message. Critics argued the presentation could make drug references look stylish or promotional.
The Times of India reported that the video was removed after online backlash and framed the controversy as an example of AI public messaging going wrong. Hong Kong’s Correctional Services Department is a government agency responsible for correctional management and rehabilitation, while broader anti-drug policy in Hong Kong is handled through official government prevention and narcotics-control channels.
Public agencies increasingly use social-media formats because young audiences may not respond to traditional pamphlets or formal notices. But the more a campaign borrows from entertainment culture, the greater the risk that style overwhelms the intended warning.
Why it matters
Generative AI can produce polished visuals quickly, but speed is not the same as judgment. A human review process still has to ask whether the target audience will receive the intended message or whether the campaign will accidentally glamorize the behavior it is trying to discourage.
The case also shows why public-sector AI use needs editorial standards. Sensitive subjects such as drugs, suicide prevention, violence, public health and youth safety require testing, expert review and plain-language safeguards. A campaign can be technologically impressive and still fail if its tone is wrong.
For governments, the reputational risk is immediate. If a campaign becomes a viral example of mixed messaging, the public may focus on the mistake instead of the safety message. That can weaken trust in future prevention efforts.
The AI governance context
The controversy shows a core problem in generative-media deployment: a campaign can meet a technical standard and still fail an editorial one. An image or video may look polished, but public communication is judged by how audiences understand it, not by how quickly it was produced.
Public agencies have additional obligations because their campaigns carry the authority of government. A private brand can withdraw an awkward advertisement and move on. A government agency working on youth safety, drug prevention or rehabilitation has to protect credibility with parents, schools, social workers and the young people it hopes to reach.
AI-generated performers add another layer. Synthetic idols can imitate the energy of popular culture without the accountability of real artists, managers or lived experience. That can be useful for cost and safety, but it can also produce tone-deaf messaging if the creative concept is not tested with the intended audience.
Good AI governance for public campaigns should include human review, subject-matter review, youth-safety review, accessibility review and a clear record of who approved publication. The more sensitive the topic, the more important it is to test whether the message could be misunderstood.
The Hong Kong case should not be read as proof that AI has no role in public-service announcements. It should be read as proof that AI content still needs the same editorial caution, audience testing and accountability as any other public campaign.
Accountability questions
The accountability question for Hong Kong officials is who approved the creative concept and what review standard applied. A sensitive public-health message should be checked not only for legality and branding but also for whether the likely audience may read it in the opposite direction.
For agencies using AI, the case shows why procurement and communications teams need shared rules. It is not enough for a vendor or internal team to produce attractive media; someone must be responsible for audience testing, harm review and final editorial judgment.
For the public, the lesson is that government AI use should be documented enough to allow criticism and correction. When a campaign is removed, officials should be able to explain what went wrong and what safeguard will prevent repetition.
How to read this story
The safest way to read the Hong Kong video controversy is as a lesson in message design, not only an AI error. Generative tools can accelerate production, but they cannot guarantee that a sensitive campaign will land properly with the public.
That means the human review chain matters. A public agency should be able to explain who approved the concept, whether the message was tested and whether health or youth-safety experts had a meaningful role before publication.
What remains unclear
It remains unclear what internal review the video received before publication, whether outside youth or health experts were consulted, and whether the department will issue revised guidance for AI-generated public materials.
It is also unclear whether the backlash will slow AI adoption by Hong Kong agencies or simply lead to stronger pre-publication review. CGN News is not asserting that AI should be excluded from public-service communication. The issue is governance.
What to watch next
Watch for statements from the Correctional Services Department, Hong Kong anti-drug authorities and public-communication specialists. Also watch whether the video is replaced with a new campaign or whether the agency publishes rules for synthetic-media use.
The practical takeaway is that AI tools do not remove responsibility from public agencies. When a message involves youth safety, public health or criminal justice, the question is not just whether the content can be generated. It is whether it should be published.
Additional Reporting By: BBC News; Times of India; Hong Kong Correctional Services Department; Hong Kong Narcotics Division