Study: AI image generators accept 85 percent of election manipulation prompts
Prominent image-based generative AI tools can be used to generate fake evidence in support of mis- and disinformation about elections, according to researchers. The study by UK-based disinformation company Logically found that Midjourney , DALL-E 2 , and Stable Diffusion accepted over 85 percent of prompts seeking to generate fake evidence that would support false claims. In one instance, p rompts relating to claims of a 'stolen election' generated images of people appearing to stuff election ballot boxes on all three platforms. Logically also was able to generate false evidence of phony claims related to elections in the UK and India. The findings raise concerns about the apparent ease with which Midjourney , DALL-E 2 , and Stable Diffusion may be used to interfere in political elections, the governance and safety of these systems, despite the acknowledgment of senior leaders at OpenAI that electoral interference is a major risk. System 🤖 DALL-E 2 Midjourney Stable Diffusion Operator: Midjourney; OpenAI; Stability AI Developer: Midjourney; OpenAI; Stability AI Country: US; UK; India Sector: Politics Purpose: Generate image Technology: Text-to-image; Generative adversarial network (GAN); Neural network; Deep learning; Machine learning Issue: Mis/disinformation
- Date it happened
- 2023-07-01
- Organisation involved
- Midjourney; OpenAI; Stability AI
- Product, system or model
- Midjourney, DALL-E 2, Stable Diffusion
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.