Since the emergence of generative AI, capable of creating realistic images from a few textual instructions, many have warned of the dangerous potential of these tools to generate hoaxes and fake news.
This is precisely what comedian Justin T. Brown wanted to expose when he used Midjourney to create a series of photorealistic images of various political figures cheating on their partners.
Thus, personalities such as Donald Trump, Barack Obama or Joe Biden are depicted in hotel rooms having extramarital affairs. All, of course, created from guide messages.
According to Mr. Brown, this series of images, published under the suggestive title AI will revolutionize the blackmail industry, demonstrates how easy it is to use these technologies as a weapon to harm others. He says it took him a little over an hour to create these images. “To show the true threat potential of AI, the project had to be a potential threat in itself, and to demonstrate that, there was no had nothing better than to attack the most powerful men in Washington,” Brown said in a statement published on the PetaPixel website.
The images were posted to a Reddit thread by Midjourney several months after he created them, and were never presented as real. Despite this, moderators immediately deleted the post, and soon after, Justin T Brown received a notification from Midjourney that he was banned from the platform.
https://twitter.com/JuuustinBrown/status/1662225616635305985?ref_src=twsrc%5Etfw%7Ctwcamp%5Etweetembed%7Ctwterm%5E1662225616635305985%7Ctwgr%5E42ca59baea9aa70f1e82baebc7f8d 2a9db273d7a%7Ctwcon%5Es1_&ref_url=https%3A%2F%2Fwww.photolari.com%2Fmidjourney- veta-a-un-usuario-por-crear-imagenes-de-politicos-cometiendo-infidelidades%2F
According to Mr. Brown, this raises several questions.
On the one hand, this shows that Midjourney has some control over what its users generate, but that they failed to act before the images were posted to Reddit, even though they had been created months ago and would have already been able to cause a lot of problems for their potential victims. On the other hand, we should ask ourselves what the criteria for content regulation are, who makes these decisions and on what basis.
Regardless, it is more than clear that there is an urgent need for a thorough debate on how to regulate the use of generative AI. According to a New York Times report, images generated by generative AI AI are already very present in the next presidential elections of 2024.