What Happened
In March 2024, as the US presidential election campaign intensified, Midjourney banned the generation of images depicting political figures entirely. The decision came after the platform had repeatedly been used to create convincing fake images of politicians in fabricated scenarios. Most notably, AI-generated images of Donald Trump being arrested had gone viral in March 2023, demonstrating the potential for AI-generated political disinformation.
Midjourney CEO David Holz acknowledged that the company lacked the moderation resources to review all political imagery at the scale of their platform, making a blanket ban the only viable option.
Timeline
Fake political images from Midjourney began circulating in early 2023, with the Trump arrest images being the most prominent. Throughout 2023 and early 2024, additional viral political fakes were traced to the platform. As the 2024 election cycle began, pressure mounted on AI image generators to prevent election disinformation. Midjourney implemented its ban on political figures in March 2024, blocking names and descriptions of major political figures from prompts.
Impact
The ban was both a pragmatic response and an acknowledgment of limitation. It demonstrated that AI image generation companies could not adequately moderate political content at scale without either massive investment in content moderation or blanket restrictions. The decision influenced other platforms’ approaches to political imagery during the 2024 election cycle.
The incident also contributed to broader discussions about AI’s role in elections, influencing state and federal legislative proposals for AI-generated political content disclosure requirements.
Response
Midjourney implemented keyword-based blocking of political figure names and descriptions. The company acknowledged that the approach was imperfect — users could potentially circumvent it through indirect descriptions — but stated it was the most responsible option given resource constraints. CEO David Holz was transparent about the limitations, stating that the company did not have the infrastructure for nuanced content moderation of political imagery.
Lessons Learned
Midjourney’s ban illustrated the difficult position AI companies face during election periods. Allowing unrestricted generation of political imagery creates serious disinformation risks, but banning it entirely limits legitimate artistic and satirical expression. The company’s choice to ban entirely rather than attempt nuanced moderation reflected an honest assessment of its capabilities.
The episode demonstrated that AI content moderation at scale requires investment comparable to major social media platforms — investment that many AI startups cannot afford. It also showed that the timing of AI capabilities relative to electoral cycles creates unique governance pressures that the industry must anticipate rather than react to.