AI Knowledge Decay: How Generative AI Is Undermining Corporate Productivity

Companies that rapidly adopted generative AI are now facing a phenomenon called “knowledge decay,” where low-quality AI outputs erode organizational trust and cost businesses an estimated $9 million annually in rework, according to a 2023 Harvard Business Review analysis.

What Is “Knowledge Decay” and How Is It Affecting Businesses?

Experts describe “knowledge decay” as a feedback loop where AI-generated content that appears polished but lacks substance undermines corporate decision-making. Oxford University professor Matthias Holweg and Babson College’s Thomas Davenport, authors of a June 2023 HBR article, note that employees often spend hours verifying or redoing AI-assisted work, compounding errors across teams.

“When AI produces work that looks good but contains mistakes or lacks depth, it creates a ripple effect,” Holweg said in a 2023 interview. “This degrades the quality of institutional knowledge over time.”

How Much Does “Workslop” Cost Companies?

The term “workslop”—coined by BetterUp Labs and Stanford’s Social Media Lab in a 2023 HBR article—refers to AI content that masks low-quality output. A survey of 1,150 U.S. workers found 41% encountered workslop in the prior month, with each incident requiring an average of 1 hour and 56 minutes to correct.

Using self-reported data, researchers estimated workslop costs $186 per worker monthly. For a 10,000-employee firm, this totals over $9 million annually in lost productivity, according to the study.

Why Is Trust in AI-Generated Work Crumbling?

Workers who received workslop reported significant trust issues. A 2023 survey found 53% of employees felt annoyed by such content, while 42% viewed the sender as less trustworthy. Half of respondents believed the colleague was less creative or reliable, and 33% said they’d avoid working with them again.

This erosion of trust extends to hiring processes. AI-generated resumes and job listings have led to “all-time lows” in confidence among both job seekers and recruiters, per HBR.

Has AI Delivered on Productivity Promises?

Despite billions in investment, most organizations report no measurable AI productivity gains. A 2023 MIT Media Lab report found 95% of firms saw no return on generative AI, while Goldman Sachs’ March 2023 analysis found no link between AI adoption and economy-wide productivity improvements.

“The efficiency argument for AI is being undermined by the need for human oversight,” said HBR. Companies now invest in verification processes to counter low-quality outputs, reversing the original goal of reducing labor.

What Are the Broader Implications for AI Adoption?

The knowledge decay issue highlights a gap between AI’s hype and reality. While proprietary models trained on company-specific data show promise, public large language models often produce “generic prose with mistakes,” according to HBR.

Workers are pushing back: A 2023 survey found 29% of employees actively sabotage AI strategies, with 44% of Gen Z workers doing so due to fears of job displacement. Meanwhile, tech sector layoffs in 2023 saw nearly half attributed to AI, despite questions about the technology’s readiness to replace human roles.

What’s Next for AI Regulation and Corporate Strategy?

Experts warn that without quality controls, AI adoption risks worsening organizational inefficiencies. Holweg and Davenport advocate for targeted AI use rather than broad mandates, emphasizing the need for human oversight.

“The question isn’t just whether AI speeds up tasks, but whether it improves decision-making,” Davenport said. “For many companies, the answer is a clear ‘no.’”