Although artificial intelligence models are exciting, they can cause some negativities. A few days ago, an image of the Pope wearing a puffer coat appeared and soon became popular on social media. This image, which is thought to be real by many people, was actually created by the artificial intelligence tool Midjourney.
This wasn’t the only artificial intelligence image that went viral lately. Images showing the arrest of former US President Donald Trump, also created by Midjourney, also spread rapidly on social media. These images, which seemed to be real, naturally caused great concern and revealed that we can understand whether the images we see on the internet are real in the future. While the discussions were going on, a move towards this came from Midjourney.
Midjourney’s free trial rights suspended for abuse
David Holz, the founder and CEO of Midjourney, announced that they have stopped the free trial of artificial intelligence with an announcement he made recently. Holz cited extraordinary demand and abuse as the reason.
Midjourney previously allowed anyone to create 25 free images. When this trial period was over, it was necessary to switch to a monthly subscription of 10 dollars. However, these rights have recently been removed due to fake images that lead to disinformation.
Holz previously said that Midjourney cannot be used to make deepfakes. However, version 5, which arrived on March 15, provided an incredible improvement in the quality of the photographs and the creation of truly indistinguishable human depictions. After that, fake images that were thought to be real began to appear on social media.
The measures taken are not enough to prevent such fake images.
Midjourney, which does not have very restrictive policies for its content, banned some words it determined “according to complaints from different countries”. However, the list containing all of these words has never been shared. Following the spread of fake images, this list was expanded by adding words such as “arrested” (image of Donald Trump).
However, this cannot be said to have caused much of a change. Trials by The Verge showed similar results when using things like “handcuffed” instead of “arrested”. This revealed that there may be a need for tighter restrictions.
In short, although artificial intelligence tools are fascinating, they also bring huge problems. As such technologies continue to develop very rapidly, we will see how companies will take steps.
35754
