Emmanuel Macron struggling with CRS, Donald Trump fleeing the police… In recent days, you must have seen funny pictures on social networks featuring public figures in crazy situations. Although very similar, these images are fake, entirely created by tools using artificial intelligence.
The ChatGPT program, at the origin of this technology, is more and more efficient. If the AI is able to hold a conversation or respond to essay topics, it can also generate images through software such as Dall-E, Starry-Ai or Midjourney.
Accessible since March 15, the latest version of Midjourney is notably capable of impressive results. From a simple sentence, Midjourney generates very realistic images. When these include public figures, one can then wonder about the risks of manipulation of information.
The future of such technology raises questions. If the rendering of these images is not yet perfect, what will it be in a few years? On social networks, the proliferation of false images is to be feared, particularly to relay erroneous information. This will “become the nightmare of fact-checking”, explain observers, according to TF1.
According to Latest news from Alsace, social networks do not yet censor these images but have already taken measures to regulate their use. Twitter, for example, can place a warning on “manipulated media” so as not to mislead its users.
The AI platforms offering these image generations have also decided to regulate their use. Dall-E, for example, refuses to generate images from the word “Emmanuel Macron”. Midjourney has banned the account of the journalist behind the fake photos of Donald Trump’s arrest.
Check out our slideshow below for some of the most impressive AI-generated images.
