Published • loading... • Updated
LinkedIn’s Hari Srinivasan Outlines New Measures to Curb AI Slop, Automated Activity
Users can flag machine-written posts as LinkedIn ramps up new classifiers and says 41% of long-form posts and 30% of short ones may be AI-generated.
- On Thursday, LinkedIn introduced a new "Seems like AI slop" button under the three-dot menu, allowing users to report content they suspect was written by machines rather than humans.
- Pangram data from early July estimated 41% of long-form and 30% of short-form LinkedIn posts are AI-generated, identifying the platform as the most AI-saturated social network.
- Chief Product Officer Hari Srinivasan confirmed LinkedIn is replacing its "enhance with AI" feature with a proofreader that does not change a user's voice, while testing private nudges for inauthentic content.
- Flagged posts will see reduced algorithmic reach similar to the "not interested" feature, as LinkedIn uses reports to tune its AI models directly.
- Across the industry, platforms including Reddit and Substack are implementing AI detection tools as automated traffic on the internet now exceeds human-generated requests.
Insights by Ground AI
Podcasts & Opinions
38 Articles
38 Articles
LinkedIn’s Hari Srinivasan outlines new measures to curb AI slop, automated activity
LinkedIn is strengthening automation defenses and detecting many automated comment attempts daily. The platform is deploying improved classifiers to identify slop and low quality content. Member feedback mechanisms will expand to flag AI-generated or inauthentic posts. LinkedIn will also test features privately alerting creators about content perception. Profile and page verification is expanding to enhance user trust and transparency.
·Mumbai, India
Read Full Article
Anyone who suspects that AI is "overboarding" behind postings can now make this known. The offer to automatically optimize texts using AI removes the platform
·Vienna, Austria
Read Full Article
Coverage Details
Total News Sources38
Leaning Left14Leaning Right3Center10Last UpdatedBias Distribution52% Left
Bias Distribution
- 52% of the sources lean Left
52% Left
L 52%
C 37%
11%
Factuality
To view factuality data please Upgrade to Premium
