A Hong Kong government anti-drug campaign meant to discourage drug use has backfired after an AI-generated music video drew criticism for glamorising narcotics before delivering its warning message. The promotional video, released by the Hong Kong Correctional Services Department (CSD) to mark the International Day Against Drug Abuse and Illicit Trafficking, was pulled following public complaints that it made illegal drugs appear attractive rather than dangerous.
The controversy has reignited debate over the use of generative AI in public awareness campaigns and the challenges of crafting effective health messaging without unintentionally promoting the behavior it seeks to prevent.
TL;DR
- Hong Kong’s Correctional Services Department released an AI-generated anti-drug music video.
- The ad featured fictional AI pop stars representing different drugs, including cocaine, cannabis, and crystal meth.
- Critics said the video highlighted the supposed “benefits” of drugs before revealing their harmful consequences, making drug use appear appealing.
- The government removed the video after online backlash.
- A second AI-generated anti-drug video was also withdrawn after it was found to contain an apparent messaging error.
- The incident has sparked wider discussion about AI-generated public service campaigns and effective drug prevention messaging.
Why did Hong Kong’s AI anti-drug ad spark controversy?
The campaign was designed to warn young people about the dangers of illegal drug use through an eye-catching AI-generated music video.
Instead of beginning with health risks, however, the video introduced a fictional pop group called Obsession, whose members personified different drugs and enthusiastically described their supposed effects.
The colourful visuals, upbeat music, and polished K-pop-inspired aesthetic quickly became the focus of criticism, with many viewers arguing that the campaign unintentionally made drugs appear fashionable before introducing its anti-drug message.
Public health experts have long cautioned that prevention campaigns should avoid inadvertently glamorizing risky behavior, particularly when targeting younger audiences.
What happens in the video?
The AI-generated video features four fictional performers, each representing a different drug:
- Icy – representing crystal meth.
- Coke – representing cocaine.
- Weedy – representing cannabis.
- Little E – representing etomidate, sometimes referred to colloquially as “space oil.”
Each character introduces themselves while singing lyrics that describe the perceived effects of their respective drug.
Only later in the video does the tone change dramatically.
The performers transform into frail, elderly figures before scenes shift to imagery showing drug paraphernalia, insects crawling across the screen, and prison scenes intended to illustrate the consequences of drug abuse.
The video concludes with a warning stating:
“Drugs are extremely harmful and can ruin a life.”
Why did viewers object?
Much of the criticism centered on the structure of the campaign rather than its overall objective.
Viewers argued that
- The video spent considerable time highlighting the supposed appeal of drugs.
- Bright visuals and polished character designs made the fictional performers look glamorous.
- The anti-drug message appeared too late.
- Younger audiences might remember the catchy songs more than the cautionary ending.
One social media user reportedly joked:
“After watching it, I want to try a few bites of each.”
While intended humorously, comments like these reflected broader concerns that the campaign had produced the opposite of its intended effect.
The government removed the campaign
Following online criticism, the Hong Kong Correctional Services Department withdrew the advertisement.
The department did not continue promoting the campaign after complaints suggested the video risked encouraging curiosity about drugs instead of discouraging their use.
The controversy illustrates one of the central challenges facing public information campaigns: balancing attention-grabbing creativity with clear, responsible messaging.
A second AI-generated ad also faced problems
The backlash did not end with the first video.
The Correctional Services Department reportedly released another AI-generated anti-drug advertisement that contained an apparent messaging error.
According to reports, the narration appeared to state:
“Whether you take or sell drugs, you won’t go to jail.”
The statement contradicted the campaign’s intended purpose and was widely interpreted as the result of an AI generation or production error.
The department subsequently removed the second video on June 27 and issued an apology.
While officials did not publicly explain how the mistake occurred, the incident has raised questions about quality control and human oversight in AI-assisted government communications.
Why is this incident significant?
The controversy extends beyond a single advertising campaign.
Governments worldwide are increasingly experimenting with generative AI to create:
- Public service announcements.
- Educational videos.
- Health campaigns.
- Social media content.
- Emergency awareness materials.
The Hong Kong case demonstrates that while AI can rapidly generate visually engaging content, effective public messaging still requires careful editorial review and audience testing.
Communications experts often emphasize that campaigns aimed at changing behavior should avoid portraying harmful activities in ways that could increase curiosity or normalize them.
What can public agencies learn?
The incident highlights several lessons for organizations using AI-generated content:
Test campaigns with target audiences
Focus groups and user testing can identify unintended interpretations before a campaign launches.
Prioritize the core message
Health and safety campaigns should communicate risks clearly from the outset rather than relying on a dramatic reveal.
Maintain human oversight
AI-generated scripts, visuals, and voiceovers should undergo multiple rounds of human review to catch factual or messaging errors.
Consider audience psychology
Behavioral research suggests that emphasizing the perceived benefits of harmful behavior, even briefly, can sometimes undermine prevention efforts.
The broader debate over AI in public communication
As governments adopt generative AI tools, experts are calling for stronger editorial safeguards.
While AI can reduce production costs and accelerate content creation, it cannot replace human judgment when dealing with sensitive subjects such as:
- Drug prevention.
- Public health.
- Mental health.
- Crime prevention.
- Emergency preparedness.
The Hong Kong campaign serves as an early example of how innovative technology can produce unintended consequences when creative execution overshadows the intended public message.
