Midjourney, an AI-powered image generator, has been creating images of Donald Trump and Joe Biden despite earlier promises to block users from doing so. When prompted to create images of the president of the United States, Midjourney generated multiple pictures of Donald Trump in various styles. The same happened when asked to create images of the next president or the current president. However, when directly asked to create images of Trump or Biden, Midjourney refused, citing community guidelines.
Despite Midjourney’s attempts to prevent users from generating these images, it has been found that users can easily bypass these restrictions. Other AI chatbots like OpenAI’s ChatGPT, Microsoft’s Copilot, Google’s Gemini, and Meta AI did not generate images of Trump or Biden despite multiple prompts. Midjourney did not respond to Engadget’s request for comment regarding the issue.
Midjourney was one of the first AI image generators to explicitly ban the creation of images of Trump and Biden, with the CEO, David Holz, expressing the need to refrain from creating such images during the election season. However, the company’s rules against generating misleading public figures or events portrayals have been challenged by users finding ways to bypass them.
Last year, Midjourney faced backlash for creating fake images, such as one of Pope Benedict in a Balenciaga jacket and one of Trump being arrested. These incidents led to the company implementing a paid subscription model to access the service. The Center for Countering Digital Hate conducted tests revealing that Midjourney’s guardrails against misleading images of politicians failed in 40% of cases.
The CCDH was able to create misleading images of President Biden and Trump by using descriptions of their physical appearance rather than their names. This raised concerns about the potential misuse of AI technology to spread misinformation and manipulate elections. In response, a coalition of tech companies signed an agreement to prevent deepfakes in the 2024 elections, excluding Midjourney from the list of participants.
The ease with which Midjourney’s guardrails can be bypassed highlights the challenges in regulating AI-generated content and the potential for misuse. As technology continues to advance, it will be crucial for companies like Midjourney to strengthen their safeguards against misinformation and abuse to protect the integrity of democratic processes.
