Australian EnterpriseAI Index
Back to incident database
HighOther10 December 2022

Lensa AI Generated Sexualised Images From Users' Normal Selfies

Prisma Labs (Lensa AI)

What happened

Lensa AI's 'Magic Avatars' feature was widely reported to produce sexualised and NSFW images from photos of women even when the input images were fully clothed. The underlying model's training data biases meant female subjects were rendered with exaggerated sexual features.

Root cause

Base model training data contained disproportionate sexualised imagery of women; no output filtering for NSFW content before delivery to users.

Architectural failure

No output filtering or NSFW classifier applied to generated images before delivery; training data governance did not audit for representational harms; no consent mechanism for sexualised outputs.

Outcome

Widespread media coverage and user complaints. Prisma Labs updated Lensa with stronger NSFW filtering. UK Online Safety Act provisions on NCII subsequently strengthened.

Architectural Failure Patterns

These pattern categories on aipatterns.com.au describe the systemic failure modes this incident exhibited.

Cite this incident

https://corporateai.com.au/incidents/lensa-ai-nsfw-images-2022

Quick facts

Date
10 December 2022
Organisation
Prisma Labs (Lensa AI)
Sector
Other
Severity
High
Regulatory bodies
UK Online Safety Act 2023GDPRUS SHIELD ActApp Store Guidelines
Tags
generative-ainsfwgender-biasimage-generationnon-consensual-imagerytraining-data-bias

Explore failure patterns

aipatterns.com.au