The Hidden Tax on Your Time: Wading Through AI Slop

For the past year, tech experts have warned that generative AI would eventually flood the internet with low-quality, automated content. The comforting assumption for many was that this digital "slop" would remain confined to obscure SEO farms and shady spam websites that actual humans rarely visit. However, new browsing data suggests a much more intrusive reality: AI-generated content is already dominating the social media feeds we scroll through every day.

According to a recent study by Pangram, a company specializing in AI text detection, a staggering 41 percent of long-form written content that users encounter on LinkedIn is likely fully generated by artificial intelligence. The situation is similar on X (formerly Twitter), where a quarter of longer articles are entirely AI-written, with another 23 percent utilizing AI assistance.

What makes this data particularly revealing is its methodology. Rather than scraping the entire internet, Pangram analyzed roughly one million posts that its Chrome extension users organically scrolled past over a two-month period. This means the numbers reflect what actual humans are viewing, reading, and interacting with, not just what exists in the digital void.

The findings highlight a fascinating behavioral paradox regarding how we use these new tools. Counterintuitively, people are overwhelmingly willing to let AI speak on their behalf in professional, real-identity settings like LinkedIn. Yet, on casual and anonymous platforms like Reddit, users are much more likely to write their own posts—AI generation on longer Reddit and Substack posts hovers at just around 10 percent. The data also notes that top-level posts on these platforms are far more likely to be outsourced to AI than the conversational comments underneath them.

Because AI models are notoriously verbose, they have become the perfect tool for users looking to quickly generate lengthy, authoritative-sounding posts. But this convenience comes with a hidden cost. Max Spero, the CEO of Pangram, aptly describes this phenomenon as a "tax on readers' time." Users are increasingly bearing the cognitive weight of constantly evaluating whether a thoughtful essay or a professional milestone was actually penned by a machine.

Social media platforms are beginning to grapple with this shift. LinkedIn, which previously integrated AI writing assistants directly into its posting interface, has recently started walking back these features in an effort to encourage authentic conversations. Reddit, on the other hand, relies heavily on its human moderators, many of whom hold strong anti-AI stances and actively delete automated spam to protect community integrity.

As artificial intelligence continues to blur the lines of digital communication, the internet is undergoing a fundamental transformation. Navigating this new landscape will require not just better detection tools, but a conscious decision about what kind of content we choose to value and engage with.

Key Points

• Pangram data shows up to 41% of long-form LinkedIn posts and nearly half of X articles are fully or partially AI-generated.
• Users are surprisingly more likely to use AI to write for their professional, real-name profiles than for anonymous accounts.
• The influx of verbose AI content creates a "tax on readers' time," increasing the cognitive load required to navigate social feeds.

Why It Matters

As AI-generated content becomes indistinguishable from human writing, users face a growing cognitive burden to verify the authenticity of the information they consume daily, threatening the core of digital social interaction.

Sources:

• LinkedIn and X Are Flooded With AI Spam, Browsing Data Suggests — 404 Media