The British Tabloid The sun Was he fooled by the photo of a flying saucer generated by an artificial intelligence? The truth is elsewhere. But where ?
“An incredibly clear photo shows a 15-meter flying saucer emerging from a storm before disappearing into space. »
This grandiloquent headline and the accompanying article were published on the website of the SunMonday September 5th. Based on a report broadcast on Mexican national television, the paper explains how a mechanic named Juan Manuel Sanchez allegedly took this extraordinary photo.
In a context where the US Navy is declassifying its archives concerning UFOs in full flight, the news has something to excite ufologists around the world. Unfortunately, for lovers of little green men, a few hours after its broadcast, the information lost its credibility.
The photo would have been created on Midjourney…
Shortly after, a Facebook user with the pseudonym of Poncho posted on the image generation group by IA Midjourney a post in which he claimed to be at the origin of the famous photographs. “Again,” he said, “looks like one of my random posts on the Facebook group has gone out of control… And now it’s on the Sun. People need to hire me to create fake news. In the light of this post, one would think that the Sun was fooled by an image created by an AIartist. But, new twist…, things seem more complicated. According to an article by Vicethe photos posted by Poncho simply have nothing to do with the images published by the Sun. Poncho has since deleted his post. Back to square one.
Will AI-generated images drive us crazy?
If the mystery of this flying saucer remains whole for the moment, this story sheds light on the doubt that the images tampered with by artificial intelligence will pose. How can we now trust images when it has become so easy to create them? In the context where fake news is already used as a weapon of digital and cultural warfare, we feel there is a new breach. For Benj Edwards, a journalist specializing in machine learning and AI, there will inevitably come a time when unscrupulous media about verifying information will be fooled by a photo generated by Midjourney or Stable Diffusion: “I think the biggest threat of AI in the short term is the destruction of our trust in communication systems. How to verify that a distant event really happened when it can be made with an AI in writing, video or photo)? “, he writes.
In fact, videos of deepfakes like that of the fake Tom Cruise already made us doubt our reality.
Now it is quite possible to create fake photos of historical figures or erotic images with the face of any celebrity in seconds. Easier to use than deepfake technology, these images will again invade the public space and produce new ethical and legal puzzles in the months to come.