ASA warns on AI-generated content and deepfakes - Lewis Silkin
The ASA has issued warnings against AI-generated deepfakes following an incident where an app used a synthetic video of singer Sabrina Carpenter to promote content
- Incident date
- Apr 2025
- Target
- Sabrina Carpenter
In April 2025, the Advertising Standards Authority (ASA) identified a problematic mobile advertisement that utilized AI-generated imagery to depict singer Sabrina Carpenter in an intimate context. The ruling serves as a stark reminder that advertisers remain fully accountable for all content, regardless of whether it is created or placed through automated AI tools.
What happened
An advertisement for an AI video generation app appeared within a mobile game, featuring an image of a man labeled "me" placed next to an image of Sabrina Carpenter labeled "My crush." The ad included an AI-generated video showing the two figures kissing, accompanied by text prompting users to "kiss your celebrity crush." A voiceover in the ad encouraged viewers to upload their own photos to generate similar intimate videos with a celebrity of their choice.
The ASA concluded that the ad was both harmful and offensive. The authority determined that the content reinforced the notion that it is acceptable to objectify women through technology by encouraging the creation of intimate deepfakes. This ruling reinforces the ASA’s stance that AI-generated depictions of real individuals must not mislead consumers into believing there is a genuine endorsement, and that advertisers must ensure all outputs are socially responsible and non-discriminatory. The ASA maintains that the use of automated creation tools does not shift the burden of responsibility; advertisers are primary stakeholders in ensuring compliance with the CAP Code.