Justin Oberman Explains LinkedIn’s ‘AI Slop’ Button and the Real Audience Detector

In a recent LinkedIn post, Justin Oberman delves into the intricacies of LinkedIn’s approach to combating “AI slop” and offers a crucial perspective on what truly matters for content creators on the platform. Oberman clarifies that LinkedIn’s algorithm has been actively suppressing content perceived as AI-generated for months, a process that often leads to a sudden drop in a post’s performance, even after a strong initial showing.

He highlights that LinkedIn’s Vice President of Product, Laura Lorenzetti, has defined “AI slop” not by its method of creation but by its intrinsic value. Oberman points out the platform’s claim of a 94% accuracy rate for its AI detection, a figure he finds astonishing but ultimately insufficient.

“LinkedIn has been battling AI slop for several months now behind the scenes with their algorithm. Instead of just pointing out whether or not a post is AI slop the algorithm just suppresses by removing the post from global recommendation feeds.”

Understanding LinkedIn’s ‘AI Slop’ Button

Oberman explains that LinkedIn has recently empowered users to report posts as “AI slop,” a move he suggests is a reaction to public discourse about the platform’s struggles with such content. He breaks down the mechanics of this reporting feature, drawing on information attributed to an internal LinkedIn source.

The ‘Signal’ and Algorithmic Audit

According to Oberman, clicking the “Seems like AI slop” button does not immediately result in a penalty. Instead, it serves as a “signal” that prompts an algorithmic audit. LinkedIn then uses its own text-detection models to evaluate the flagged post.

“LinkedIn passes flagged posts through its own native text-detection models. LinkedIn asserts its internal detection engine successfully filters out false positives with a 94% accuracy rate.”

Protection Against Malicious Reporting

A key point Oberman emphasizes is LinkedIn’s built-in safeguards against malicious reporting. He notes that if a competitor flags a genuinely human-written or valuable post, the algorithm is designed to recognize the human cadence and disregard the false report.

“If a competitor flags your 100% human-written or valuable post, the algorithm is built to recognize the human cadence and override the malicious report.”

Furthermore, Oberman clarifies that a confirmed “slop” flag does not lead to account bans. The primary consequence is a limitation on the post’s reach, restricting it to the user’s immediate network. However, he cautions that a consistent pattern of low-performing posts could still impact overall reach.

The Ultimate AI Detector: Your Audience

Despite the technical measures LinkedIn employs, Oberman posits that these internal mechanisms are secondary to a more critical evaluation tool: the audience’s perception.

“Of course, none of this really matters. Because the only AI detector you should be worried about is the one inside the head of your audience.”

He argues that the true measure of content success lies in its ability to resonate with human readers. To aid creators in this endeavor, Oberman has developed “The Copy Hunter Field Guide,” a resource designed to help refine writing to sound authentically human and engaging.

In essence, Oberman’s analysis shifts the focus from platform-specific technicalities to the fundamental principles of creating valuable, human-centric content that captures and holds the attention of the intended audience.