As artificial intelligence rapidly transforms how organizations create value, a new working paper from the MIT Sloan School of Management argues that the next competitive advantage will not come from producing more with AI—but from knowing how to verify it.

According to researcher Christian Catalini, the cost of generating sophisticated outputs with AI is falling towards zero, while the human capacity to review, validate and assume responsibility for those outputs remains limited by time and biology. The result is what he calls a growing “measurability gap”: AI can increasingly produce work faster than people can assess its quality.

Rather than treating AI as a faster employee, Catalini argues that leaders should see it as a powerful but unpredictable agent that requires clear direction and human oversight. Without that balance, organizations risk creating what the research describes as a “hollow economy”—one driven by speed and volume, but weakened by unchecked errors and declining trust.

Instead, the paper outlines a practical playbook for organizations looking to build what Catalini calls an “augmented economy,” where human judgment remains central to decision-making.

The research highlights several priorities for executives:

• Make verification a strategic capability. In an AI-driven economy, trust and validation become sources of competitive advantage—not simply compliance functions.
• Adopt a three-layer operating model. Define human intent, let AI execute at scale, and ensure expert human review before critical decisions or outputs.
• Invest in high-quality verification data. Reliable datasets, particularly around failures and edge cases, will become increasingly valuable for improving AI performance.
• Rethink talent development. As entry-level work becomes more automated, organizations will need new ways to develop the next generation of experts capable of supervising intelligent systems.
• Prioritize uniquely human capabilities. Judgment, accountability, relationship-building and strategic decision-making become more valuable as routine knowledge work is increasingly automated.

Catalini also suggests that individuals should increasingly focus on learning how to direct AI systems effectively rather than simply performing tasks themselves, while companies should redesign workflows to ensure human expertise remains embedded in every critical decision.

—

Source: MIT Sloan