USA 
United Kingdom
Companies rushing to integrate generative AI are finding a puzzling disconnect with adoption rates soaring, but meaningful returns still scarce.
A study from the MIT Media Lab revealed that 95% of organizations have seen no measurable ROI on AI investments, despite AI-led processes nearly doubling last year. Use of AI in the workplace has also doubled since 2023, yet the benefits are failing to materialize.
BetterUp Labs, working with Stanford’s Social Media Lab, believes one explanation is what it calls “workslop", referring to AI-generated work that looks polished but lacks the substance to advance a task, effectively shifting the burden downstream to colleagues.
‘Invisible tax’ on productivity
Survey data from 1,150 full-time US employees found 40% received "workslop" in the past month, estimating that 15.4% of the content they see at work fits the definition. Incidents flow mostly between peers (40%), though 18% travel upward from staff to managers, and 16% flow downward from leaders to teams.
Each case of workslop consumes an average of one hour and 56 minutes to fix, representing an “invisible tax” of $186 per employee monthly. For a workforce of 10,000, it equates to more than $9 million annually in lost productivity.
Workers also report emotional strain. Over half (53%) feel annoyed by workslop, 38% confused, and 22% offended. Trust suffers too, as 42% see colleagues who send workslop as less trustworthy, 37% as less intelligent, and nearly half as less capable. A third of recipients say they are less likely to want to work with that colleague again.
One finance worker said AI-created drafts “further the agenda of creating a mentally lazy, slow-thinking society.” A tech manager reported spending hours clarifying content, while a retail director described wasting time verifying information and redoing tasks.
Turning workforce data into early warnings for high-cost employees
Many employers only learn about high-cost claims after the fact, relying on annual health plan reports that provide little opportunity for prevention. Yet when absenteeism, disability, and workers’ compensation are included, the top 5 percent of cases drive nearly 60 percent of total costs. Looking only at medical and pharmaceutical claims limits an employer’s ability to understand where risk is forming and how costs escalate over time.
By integrating medical, pharmaceutical, disability, absence, compensation, and broader human capital data, employers gain a more complete and predictive view of workforce risk. Workpartners’ Human Capital Risk Index (HUI) leverages this integrated data warehouse to flag emerging high- and moderate-risk cases early, enabling timely, HIPAA-compliant outreach and clinical prevention.
Through a holistic, person-centric care model, individuals receive high-touch support across health, work, and family dimensions—helping shorten or prevent periods of high risk and high cost. The result is earlier intervention, improved outcomes, and measurable reductions in utilization, lost time, and total cost.
What You’ll Learn
Why high-cost claims are often identified too late
How integrated data improves risk prediction
How our HUI flags emerging risk early
Why holistic, person-centric care matters
How early intervention reduces total cost
How leaders can respond
BetterUp Labs warns that indiscriminate promotion of AI use fuels indiscriminate results. “Gen AI is not appropriate for all tasks, nor can it read minds,” the report said, urging firms to set clear guidelines around best practices and tools.
Researchers also emphasized the importance of mindset. Employees with high optimism and agency (dubbed “pilots”) use AI 75% more often at work than their “passenger” peers, and are more likely to apply it purposefully to enhance creativity. Passengers, by contrast, are more prone to using AI as a shortcut to avoid work.
The study argues that collaboration must remain central. Workslop undermines teamwork by offloading cognitive effort to others, creating confusion and damaging relationships. Leaders are encouraged to model intentional AI use, establish guardrails, and frame the technology as a tool to support shared outcomes rather than a way to dodge responsibility.
Workslop may be effortless to produce but carries a heavy cost for organizations that fail to control it.
