Core Development: The gap between promised AI efficiency and actual implementation has created an unexpected management burden for organizations. Rather than streamlining work, AI deployment is requiring substantial employee effort for prompting, validation, and error correction—work that often goes uncounted and unmeasured.[1] A January 2026 Workday study found that over a third of time saved through AI is offset by rework, creating what the report calls an "AI tax on productivity," with net value (time saved minus time lost) often lower than expected.[1]
Key Players & Context: The tension involves multiple stakeholders: C-suite executives driven by FOMO to deploy AI quickly; middle managers caught between executive pressure and employee resistance; and frontline workers absorbing unplanned labor. According to Rumman Chowdhury, former U.S. Science Envoy for AI, executives are "incentivized to pretend like it works really well," while responsibility for failures falls on employees who had no input in the adoption decision.[1] Companies like ServiceNow's Kellie Romack are discovering they must manually catch AI errors (like basic math mistakes) and spend hours verifying citations and hallucinations.[1]
Why It Matters Now: Half of organizations piloted general-purpose AI tools in 2025, but adoption and readiness are mismatched.[1] The AI industry oversold capabilities—OpenAI CEO Sam Altman's claim of "PhD-level experts in your pocket" contrasts sharply with reality.[1] Meanwhile, organizations also face a paradoxical shift: while AI automates routine managerial tasks at some companies (Amazon, Moderna, McKinsey), research analyzing 375 million U.S. job postings shows AI adoption is actually increasing demand for managerial roles focused on AI strategy, integration oversight, and governance rather than reducing them.[3]