Claim
The gap between AI capability and institutional response is not a lag that will close once decision-makers catch up. It is a structural condition that widens as long as capability growth continues, because deliberative institutions are built to operate at human pace and cannot accelerate to match an exponential curve.
Mechanism
Committees, legal processes, and organizational governance are designed for stability. They update through deliberation, which takes months to years. AI capability thresholds cross unexpectedly and repeatedly. Each crossing forces reactive policy change from institutions that had just settled their previous response. The gap is not a one-time alignment problem; it regenerates at each capability threshold. Mollick's argument is that turbulence is therefore the expected state, not a transitional phase. Organizations that plan around the assumption of future stability are planning for a condition that does not arrive.
Conditions
Holds when: AI capability is on an exponential or near-exponential growth curve. Applies to any institution (corporate, regulatory, academic, government) that sets policy through deliberative processes slower than the capability curve moves.
Fails when: capability growth plateaus or slows significantly. If progress stalls, institutions have time to catch up and the structural gap closes.
Evidence
Mollick names the mechanism directly in his June 30 essay:
"The instability is what happens when institutions that move at the speed of people (or worse, committees) try to track a capability curve that is very much not human in nature."
He extends the implication to the trajectory:
"And as long as we are on some sort of exponential, and for as long as it lasts, the gap only widens."
The essay does not offer empirical data on institution response times; the argument is structural and derived from the mismatch in operating speeds.
Signals
- Your organization's AI policy is already outdated by the time it is ratified.
- Teams regularly work around policy rather than waiting for it to catch up.
- Regulatory filings describe capabilities that current models already exceed.
- "Wait until things settle down" is a recurring justification for deferring AI adoption decisions.
Counter-evidence
The structural lag argument depends on continued exponential growth. Several AI researchers argue that capability growth is already slowing in specific domains (e.g., raw benchmark performance on coding and math). If progress becomes incremental rather than step-function, institutions may have adequate time to adapt. Mollick acknowledges the exponential assumption in his qualifier ("for as long as it lasts") but does not give a timeline or probability for when it might slow.
Cross-references
- AI labs have achieved their founding goal, requiring a shift from co-intelligence to co-existence as the operating frame for knowledge work (earlier Mollick frame on the shift in how AI is used)
- Agentic AI inverts the AI-access argument: it amplifies domain experts rather than equalizing non-experts (companion card from the same essay on who benefits from agents)
- A 14-hour autonomous agent run completes 2 to 17 weeks of engineering work for $251 in token costs (companion card from the same essay on agent run economics)