AI spam dominates LinkedIn and X feeds, detection study shows

The detection firm publishing this data has a direct commercial interest in making the problem look large and measurable, so the headline numbers deserve scrutiny before being treated as neutral findings. What the summary doesn't address is whether LinkedIn and X have actually seen these reports and chosen inaction, or whether the data simply hasn't reached enforcement teams in a form they can act on.

This story sits in the same credibility-infrastructure conversation as the OpenAI SWE-Bench Pro finding covered the same day from The Decoder. In that case, a widely trusted measurement tool turned out to be unreliable, undermining how the industry validates progress. Here, the inverse problem surfaces: detection tools can now quantify synthetic content, but the measurement itself is contested and commercially motivated. Both stories point to the same structural gap, which is that the field is building capability faster than it is building trustworthy ways to audit what that capability produces or pollutes.

Watch whether LinkedIn issues any public response to the detection firm's methodology within the next 60 days. A platform rebuttal would signal that enforcement infrastructure exists but is being withheld; silence would confirm that the economic incentive to act simply isn't there yet.