AI Content Flood: Can Social Platforms Survive the Onslaught?

Artificial intelligence-generated spam has become a defining challenge for social media networks. As AI models grow increasingly sophisticated, they can now replicate human-like interactions and produce realistic images and video clips, muddying the waters of online discourse. But which platforms are feeling the worst effects of this AI deluge?

According to a recent report, LinkedIn is the hardest hit, with 41% of all long-form posts in the app showing indicators that they were created by AI tools. The data comes from Pangram, a technology company that has developed a browser extension for detecting AI-generated writing online. No AI detection software is 100% accurate, so it's difficult to gauge the full scope of the problem. Yet Pangram's tracking—based on average measurements across social media platforms—suggests LinkedIn hosts the most AI-generated content, followed by X, where roughly a third of long-form posts appear to be AI-written.

Every platform is grappling with AI spam. Pinterest has worked to address concerns about the influx of AI-generated posts, which some users say have flooded their feeds. It is the only platform so far to add an AI off switch, of sorts, giving users the option to limit the prevalence of AI-generated content.

Meanwhile, Reddit recently shared insights into its efforts to combat a rise in AI-generated comments and activity. X has also acknowledged its challenges in dealing with AI-generated content.

In March, Nikita Bier, X's head of product, announced a change to the platform's creator monetization incentives. The update was designed to address a sharp rise in AI-generated deepfakes related to the war in Iran.

Bier and X owner Elon Musk have both discussed the rising impact of AI bots on social apps in general. Musk said in a March 2023 post on X that AI development has made it impossible to keep bot armies out of social apps, because it's “trivial to spin up 100k human-like bots for less than a penny per account.” Musk's proposed solution was to charge every user for access, assuming that even a small fee would restrict the creation of bot accounts at scale.

But not enough people showed interest in paying for X. In fact, paid social options in general have never resonated with audiences.

On LinkedIn, AI countermeasures have included reach penalties when AI spam is detected.

Nonetheless, most social platforms are also heavily invested in AI projects. As such, they actually want users to integrate AI tools into their online experience, rather than switching them off. Meta, for example, has invested hundreds of billions of dollars into AI development as part of its effort to become the global leader in AI technology.

But Meta users haven't been especially enthusiastic about the AI tools Meta keeps pushing at them. These days, Facebook and Instagram users are regularly prompted to use Meta AI to ask questions, generate images of themselves, and compose posts. Yet when Instagram Chief Adam Mosseri recently announced that IG was integrating the company's latest Muse Image generation tools into the app, replies to his post were almost universally negative.

Some of the comments included “AI slop,” “We need an option to turn off AI-generated content,” and “Literally no one wants this.” These user responses reflect the rising backlash against AI tools, at least when it comes to recreational consumer use.

The response is no surprise because in many respects, AI tools are the antithesis of social media engagement. Social media was built on the concept of connection, which, at least until now, has meant human connection. Meanwhile, AI tools can only replicate that interaction via simulated personalities and replies, while AI-generated images and videos have begun replacing people's real experiences, real artwork, and real life events.

What's the point of social apps if none of what's being posted is real, and if they're no longer a means to showcase personal experiences and create connections? If the majority of what's being shared online ends up being fake, then that erodes trust in the whole network and weakens the key value proposition of social media until it vanishes.

Then social media becomes just entertainment, with platforms churning out an endless stream of videos from people no one knows, and replacing TV screens with smaller screens and shorter shows. That does seem to be the trend. In that case, what even is a social media app in the AI era?

Are messaging apps now social media? Because that's where people are engaging. For many users, that's now the only place they're engaging in real social interactions online. But that also shrinks the value of social media as a medium and moves it more towards a curated entertainment app that responds to user interests. And in that paradigm, AI-generated content probably won't win out, because the sheer amount of AI content being posted means that quality will remain an issue. For every cool video of Harry Potter characters in modern styles, there will be thousands more forgettable junk depictions with legs in weird places, over-the-top facial expressions, and misaligned lip syncs.

AI will get better, for sure, but there's a reason the majority of successful AI videos use established IP. Because those original ideas still resonate more than what most people can come up with, and empowering more creativity could eventually lead to less user interest.