AI Slop: AI images as a problem for YouTube, Instagram, and the internet

AI images and AI videos are found on all common platforms. They will change the internet. When it comes to cheaply made content primarily intended to generate attention, it's called AI Slop. But AI also at least dilutes the information landscape. It's everywhere. However: Reactance is a deeply human-rooted reaction – a psychological phenomenon where people resist when they feel their freedom or autonomy is threatened.

The fact that many people are already tired of AI content is evident not least in protests, complaints, and rejections of AI topics, content, and, for example, AI commercials like those from McDonalds and Coca-Cola recently. AI is treated like a new intern when something appears to be of inferior quality somewhere. And everyone has probably heard jokes about hyphens, AI-generated LinkedIn posts, and AI graphics that are supposed to present complex topics in a seemingly simple way.

"We won't be able to trust our eyes anymore"

Nevertheless, Instagram CEO Adam Mosseri assumes that AI-processed images will change the aesthetics of social media more than anything else. His grim prediction, which he posted in an end-of-year post: "We won't be able to trust our eyes anymore." While we have been able to assume our entire lives that images are real snapshots, according to Mosseri, this will now change. "This is going to be uncomfortable – we are genetically predisposed to trust our eyes." Mosseri's solution: Develop the best AI tools, but also secure labels for AI-generated content as well as for authentic content. He warns: "Authenticity will be infinitely reproducible."

At the same time, Mosseri declares the public sharing of personal moments to be dead. It has been for years. He means that private individuals no longer post photos from their private lives. That has shifted to direct messages. Status updates on WhatsApp or Signal have likely also contributed to this. This differentiation of platforms for various purposes has been ongoing for a while. According to this, social media only features professional content.

However, creators are now focusing on precisely this "raw style" and are not showing AI-embellished images. If perfect images are cheap to create, imperfect images will become a trend. Reactance includes people rejecting the trigger, doing the opposite of what the sender intends. Intrusive advertising leads to rejection. Paris Hilton was exciting in the 90s, but when people got tired of her, she became an object of hate.

Known sources instead of AI uncertainties

Mosseri also assumes that in the future, we will rely more on background information about an account, a sender, or a content creator in order to be able to trust. This applies to social media accounts as well as publishers and media companies. Since AI answers and summaries, whether in search engines or chatbots, can always be interspersed with false information and hallucinations, it is possible that people will increasingly rely on known sources.

Furthermore, it is not only easy to fill a social media account with AI-generated posts. Entire websites can also be created in an instant thanks to AI – complete with fake news, deep fakes, and simply cheap copies. In doubt, there are people with malicious intentions behind them who want to spread propaganda and false information.

On YouTube, every third video shown to a new user is reportedly AI Slop – meaning cheaply AI-generated videos that are usually intended to be monetized or often contain questionable information. This is shown by research by Kapwing, a platform for video and image editing, which of course also offers AI tools. The study addressed the question of how lucrative cheaply produced AI videos are. Kapwing estimates that 117 million US dollars were recently paid out to the creators of such videos.

(emw)