Humans Must To Verify AI Solutions

A new report from MIT Sloan Management Review and Boston Consulting Group underscores a critical truth about artificial intelligence (AI): responsible AI cannot succeed without human expertise.

In its fifth yearly study, an international panel of academics, executives, and policymakers overwhelmingly agreed — 84% of respondents — that AI governance fails if organizations neglect to cultivate human experts capable of verifying AI solutions.

Far from being a narrow “output check,” verification is described as a holistic process spanning the AI lifecycle. Experts argue it involves interpreting context, auditing workflows, setting thresholds, and knowing when not to rely on AI at all. “Context matters,” said Ryan Carrier, founder of ForHumanity, emphasizing that many risks are societal rather than technical, such as misalignment with public values or harm to vulnerable groups.

Panelists warned that delegating verification solely to machines erodes institutional capacity. Consultant Linda Leopold cautioned that over-reliance on AI could cause human judgment to atrophy, leaving organizations unable to challenge or oversee systems effectively. Others highlighted the dangers of “compliance theater,” where governance frameworks exist in name but lack meaningful human oversight.

Still, experts acknowledged the limits of human verification at scale. Wharton’s Kartik Hosanagar noted that exhaustive checks are infeasible for systems processing massive datasets. Instead, leaders are urged to adopt hybrid models where human judgment is strategically deployed alongside automated tools.

The report concludes with five recommendations:

• Embed human oversight throughout AI design and deployment.
• Combine human judgment with automated tools to extend scale.
• Invest in cultivating domain-savvy experts.
• Scrutinize not just outputs but organizational lessons learned.
• Treat verification as a strategic imperative, not a compliance exercise.

Without empowered human verification, responsible AI becomes theater. With it, AI becomes a true force multiplier for trustworthy, value-driven impact.

Here’s the article.