Companies racing to deploy artificial intelligence are confronting “AI debt,” the hidden costs of poorly implemented autonomous systems, according to Asana’s State of AI at Work report surveying more than 9,000 knowledge workers across the U.S., U.K., Australia, Germany, and Japan. Some 79% of firms expect to incur AI debt, which can show up as security risks, bad data, ineffective agents that waste employee time, and a widening management skills gap.
AI adoption climbed to 70% in 2025 from 52% a year earlier, but worker strain is rising: digital exhaustion reached 84% and unmanageable workloads 77%. Researchers at BetterUp Labs and Stanford’s Social Media Lab say 40% of U.S. desk workers encounter AI-generated “workslop”—polished but vacuous output—adding nearly two hours of extra work a month and an estimated $9 million annual productivity drag per company in the study.
Experts urge companies to pilot and sandbox AI tools, define clear use cases, and train staff rather than assuming instant productivity gains. Without guardrails and governance, executives risk trading speed for costly rework, lost time, and employee burnout.
Related articles:
AI at Work Is Here. Now Comes the Hard Part
AI Risk Management Framework
OECD AI Principles
