---
type: "synthesis"
theme: "governance-speed"
sources: ["governance"]
id: "cross-governance-speed-gap"
sourceVaultSlug: "hbr-seg-governance"
originDay: 7
articleStem: "hbr-seg-governance"
sourceUrl: "(unified vault: 8 sources)"
sourceTitle: "HBR — Firm Ⅱ-B · Governance, decision rights, leadership, risk"
---
# The Velocity Mismatch

The corpus's most quantitative shared claim is that **governance moves in years while AI moves in months.**

- *AI Nightmares* puts a number on it: a standard RAI policy takes a *minimum of one year* to approve while AI changes *monthly* ([[claim-standard-rai-too-slow]]) — the [[concept-agentic-ai-governance-gap]]. The [[concept-ethical-nightmare-challenge]] and its 6–10-week pilots are the speed fix.
- *Decision-Making by Consensus* argues AI *compresses decision cycles* until slow consensus becomes fatal ([[claim-consensus-fatal-post-ai]]), demanding the [[concept-wartime-disposition]].
- *Boards Are Falling Short* locates the same lag in regulation: [[claim-regulators-poorly-positioned]] and [[contrarian-regulations-lack-value]] — rules are "outdated on arrival."
- *How C-Suite Roles Are Reshaped* dates the tipping point: AI-as-board-hygiene is "table stakes by 2027."

**The counterweight the corpus supplies to itself:** speed has a floor. *Can AI Agents Be Trusted?* shows that stripping oversight to gain speed is self-defeating ([[claim-micromanagement-defeats-purpose]]), and A059's own [[question-human-in-the-loop-bottleneck]] asks whether mandating a human reintroduces the very lag it condemns. The mature synthesis across all four: *differentiated, risk-based oversight* — automate the low-risk/high-volume, keep humans on the high-stakes — rather than uniform speed or uniform control.