AI generated deepfake of Pope Francis made using Midjourney. Credit: Pablo Xavier / Guerrero Art
Midjourney has suspended free trial access to its AI (artificial intelligence) image generator following a surge of deepfakes going viral on the internet.
Midjourney allows users to create AI generated images by entering textual prompts. Midjourney and other platforms like it have exploded in popularity, but there are ethical concerns that AI image generators could be used to create misleading deepfakes of politicians, celebrities, and public figures.
Deepfakes use a form of AI called deep learning to create images or videos of fake events. Although deepfakes can be used for entirely benign entertainment purposes, there are growing worries that they could be used to engineer deceitful narratives, particularly in relation to politics or public figures.
Midjourney shuts off access to free trial
“Due to a combination of extraordinary demand and trial abuse we are temporarily disabling free trials until we have our next improvements to the system deployed,” commented David Holz, the founder of Midjourney, on the company’s Discord server.
In a follow-up post on Wednesday, Holz added “We tried turning trials back on again with new safeties for abuse but they didn’t seem to be sufficient. We are turning it back off again.”
The Washington Post reported that the company’s decision to halt access to its free trial was due to a surge in the creation of deepfakes on the platform, with some notably depicting figures like former US President Donald Trump and Pope Francis.
However, the company’s CEO has denied this. “There were some pretty big misunderstandings in the [Washington Post] article,” Holz told Gizmodo. “We stopped trials because of massive amounts of people making throwaway accounts to get free images…This happened at the same time as a temporary gpu shortage. The two things came together and it was bringing down the service for paid users.”
Holz added that the company would like to resume free trials for Midjourney users but they are still trying to work out a suitable way to do this.
Deepfake image of the Pope shared on Reddit. Credit: Pablo Xavier / Reddit
Rise of deepfakes
The main concern with the rising prevalence of deepfakes is that they might be used to create and spread misleading news. This could be especially problematic in countries experiencing deepening political and social polarization.
Deepfakes can be very convincing, as demonstrated this week when an image of the Pope wearing a Balenciaga puffer jacket and a large silver crucifix went viral on the internet. Many online users were genuinely fooled by the image, which first appeared on Reddit last Friday.
The creator of the deepfake has since been suspended from the site but he did provide some of his thoughts on the viral image to BuzzFeed.
“I’m trying to figure out ways to make something funny because that’s what I usually try to do,” Pablo Xavier told BuzzFeed News. “I try to do funny stuff or trippy art — psychedelic stuff. It just dawned on me: I should do the Pope. Then it was just coming like water: ‘The Pope in Balenciaga puffy coat, Moncler, walking the streets of Rome, Paris,’ stuff like that.”
Pablo Xavier, who declined to give his surname in order to protect his privacy, said that he used Midjounrey to generate the deepfake. He was surprised at how quickly the image went viral.
Although the AI generated image of the Pope was meant as a light-hearted joke, some online users took offence to the picture, whereas others thought it was real and criticized the Vatican for lavish spending.
Source: Greek Reporter
The views expressed in this article are the author’s own and do not necessarily reflect Coverpage’s editorial stance