Midjourney, a frontrunner in AI image-generation, is discussing a full prohibition on AI-generated pictures of Joe Biden and Donald Trump as the 2024 US elections approach. The move seeks to prevent the platform from being used as a medium for disseminating false information ahead of the polls.
David Holz, CEO of Midjourney, openly expressed his reservations regarding political content on his platform, declaring his uncertainty regarding the importance of political speech on the platform in the coming year. He also humorously conveyed his ambivalence towards generating images of Donald Trump, despite finding him aesthetically captivating.
The controversy over AI-generated graphics, particularly in political scenarios, has escalated following scandals involving illicit AI-generated images of Taylor Swift. Taylor Swift was targeted in a campaign of counterfeit images spread across social media, garnering millions of views.
The influence of AI on political dialogue continues to grow, raising concerns about its ability to concretely interfere with democratic procedures. There have been numerous instances of politically motivated deep fakes, often utilized to stir up controversy or propagate misinformation. In a shocking recent case, an AI avatar of Pakistan’s ex-Prime Minister claimed an election success prior to the final tally.
CEO of OpenAI, Sam Altman, expressed concerns over the new capabilities of AI leading to improved deep fakes compared to those in 2020. In response, OpenAI has begun embedding unique metadata into images produced with DALL-E 3 to help platforms and users identify AI content.
Meta has also stepped up its efforts to detect AI-generated content across its platforms. However, no single strategy has proven to be entirely foolproof to date. Regarding AI-generated images of Trump or Biden, Midjourney, for the time being, is considering steering clear of it.
