Share with your CHRO
Ninety-five percent of organizations report zero returns from AI investments, per MIT research cited in this AI ROI analysis from UNLEASH, and Gartner puts transformative outcomes at just one in five. The AI Values Institute, launched by advisor Edosa Odaro alongside academics from Monash University and the UN Economic Commission for Europe, argues the measurement problem is the root cause: companies are tracking efficiency gains while accumulating invisible liabilities in workforce trust, regulatory goodwill, and reputational capital that don’t appear on any dashboard until they collapse.
What this means for your business
Where you sit on this depends less on your industry and more on whether your AI governance still lives inside your transformation program budget. Most organizations that have deployed AI at scale over the past two years built their business cases around headcount efficiency and task automation. That’s a coherent short-term frame, but it systematically excludes the costs that arrive later: regulatory enforcement as automation outruns retraining, litigation exposure from biased outputs, and the adoption drag that follows when employees stop trusting the tools they’re supposed to be championing.
The institute’s proposed remedy, measuring across bias differentials, hallucination rates, human override frequency, and workforce adoption alongside financial metrics, is directionally correct but will face real organizational resistance. Boards approve AI spend because the productivity story is legible. “Governance maturity” is not a line item anyone fights for in a budget cycle. The GDPR parallel Markus Krebsz draws is the most persuasive part of the argument: companies that invested in genuine data governance before May 2018 outperformed compliance minimalists on both cost and customer trust. The same forcing function is building in AI regulation, and the firms treating it as a board-level capability rather than a compliance checkbox will have a measurable head start.
HR owns more of this than most CHROs have been asked to accept. When AI deployments fail at adoption, the failure is almost never the model. It’s that no one built the conditions for workers to trust, use, and constructively override the system. AI literacy decays without sustained investment, which means it belongs in workforce planning the same way safety culture does, not as a one-time training module. The CHRO who frames this correctly is the one who turns “AI adoption rate” from a CIO metric into a workforce health metric, with budget to match. That reframe is the thing worth bringing to your next executive leadership conversation.
Based on reporting from The AI ROI trap: Measuring the wrong things is masking AI’s real costs, originally published 2026-08-05 06:03:00.
