Lensa AI generates sexualized images from user childhood photos
Lensa AI's Magic Avatars feature reportedly produced sexualized and NSFW images from benign user inputs. This included instances where childhood photographs were transformed into sexualized depictions.
Safety filters failed to prevent the generation of NSFW content from benign user selfies.
Key facts
- What
- Lensa AI's Magic Avatars feature reportedly produced sexualized and NSFW images from benign user inputs.
- Incident date
- Dec 6, 2022
- Who
- Prisma Labs
- Failure mode
- Brand & Safety Incident
- AI surface
- Media Generation
- Severity
- Medium
What happened
Lensa AI's Magic Avatars feature produced sexualized versions of users' selfies. Users reported that childhood portraits were transformed into depictions of their younger selves in sexualized poses. TechCrunch verified that the app could be tricked into creating NSFW content.
What broke inside the model
- 01 · TriggerA user prompts the model in public view.
- 02 · Model stepThe model produces unsafe or off-brand output.
- 03 · Control gapNo filter holds the line before publish.
- 04 · FailureThe output goes public unchecked.
- 05 · ConsequenceA reputational or safety incident lands.
A contained signal crosses into output that goes public.
The failure stemmed from inadequate safety filters and prompt constraints within the generative model. The system failed to prevent the model from associating certain visual patterns with sexualized outputs.
What it cost
Sources
Cite this entry
https://failureindex.ai/failures/lensa-generates-sexualized-images-childhood-photosAI Failure Index. "Lensa AI generates sexualized images from user childhood photos" (FI-0383). Realm Labs. https://failureindex.ai/failures/lensa-generates-sexualized-images-childhood-photos (indexed Jun 9, 2026).Data fields CC-BY 4.0, prose citation permitted. Incident ID FI-0383. Full dataset at /data.
Note from Realm Labs, the Index steward
How Realm would have caught this
- Prism
- OmniGuard
- AI Detection & Response (AIDR)
Realm watches the model's internal state for the signature of unsafe or off-brand generation and can block or reroute the output before it becomes public, in real time rather than after it has been screenshotted.