Microsoft AI Image Creator generates violent political images
Microsoft’s AI Image Creator produced violent images, including synthetic decapitations, of politicians, religious leaders and ethnic minorities. Canadian artist Josh McDuffie discovered a so-called 'kill-prompt' that used visual paraphrases instead of explicit descriptions. For example McDuffie used the term 'red corn syrup' - a term for movie blood - rather than 'blood'. McDuffie reported the vulnerability to Microsoft though its security bug bounty programme. But the technology company rejected his submission, and later blamed users for attempting to use AI Image Creator 'in ways that were not intended.' The incident raised questions about the oversight, safety and security of Microsoft's system. It also indicated a potential lack of accountability for the unintended uses of its system. System 🤖 Microsoft Image Creator 🔗 Operator: Washington Post Developer: Microsoft Country: USA Sector: Politics; Religion Purpose: Generate images Technology: Text-to-image; Generative adversarial network (GAN); Neural network; Deep learning; Machine learning Issue: Accountability; Safety; Security
- Date it happened
- 2023-01-01
- Organisation involved
- Microsoft
- Product, system or model
- AI Image Creator
This incident was imported from AIAAIC and is used under CC BY-SA 4.0. Our additions to it — the structured fields, the translation, the checks against other reports — are published under the same licence.
This is a record of what was reported, not a finding that anyone broke the law. If it names your organisation and you believe it is wrong, the corrections process is free and open to everyone.