Microsoft Exec Warns: Low-Quality AI Content Is Creeping Into Daily Work

Microsoft executive vice president Ryan Roslansky recently posted on LinkedIn, warning that low-quality AI-generated content has spread from social networks into daily office work, creating a vicious cycle of "AI writes, AI reads, AI summarizes." He mentioned receiving a document that was clearly AI-made, and summarizing it with AI brought no new insights—the value just got diluted.

Roslansky believes generative AI can speed up drafts and summaries, but it can't guarantee accuracy or logical completeness. He suggests that humans should handle key decisions, data checks, and forming viewpoints, and that teams should set up rules for accountability and source verification.

To tackle this, LinkedIn launched a "Seems like AI slop" feedback option on July 30, and it's using classifiers to spot and reduce low-quality AI content in users' feeds.