With the rise of AI tools capable of rapidly creating altered images and videos, the spread of political misinformation is becoming easier in the lead-up to the US presidential election. In response, Midjourney is considering a ban on political images, as reported by Bloomberg. CEO David Holz mentioned during a Discord chat that the company may prohibit images of figures like Biden and Trump over the next year. “I know it’s fun to make Trump pictures — I make Trump pictures,” he stated. "However, it might be better to take a step back during this election.”
Previously, Midjourney's AI was utilized to generate deepfakes, including widely circulated images of Trump being arrested and the pope in a Balenciaga-inspired coat. In light of this, the company ended free trials for its AI image generator. Currently, Midjourney has policies against creating “misleading public figures” and portrayals of events that could cause confusion. While Bloomberg managed to create a humorous image of Trump with spaghetti using an older version of the system, the updated version refused similar requests involving the former President.
Even with a ban on high-profile political images, Midjourney may only be shielding itself from criticism, without curbing the overall use of AI in political disinformation. Other tech companies are also taking measures to combat misinformation. For instance, ChatGPT is set to begin tagging images generated with DALL-E 3, while Meta is developing technology to identify AI-created images, videos, and audio clips.