Slow Down to Speed Up: Jonathan Spector on Establishing AI Guard Rails
SummaryJonathan Spector, Founder & Principal Advisor at Converge Strategic Partners, joins the podcast to unpack how emergency managers — especially solo practitioners and multi-hatters — can adopt AI deliberately and safely. Drawing on his recent All Hazards Consortium (AHC) webinar, Jonathan walks through the minimum governance decisions needed before touching AI, how to verify outputs when the clock is running, how to build a lightweight audit trail, and how to avoid the trap of passive AI dependency. The conversation closes with a sharp look at "trendslop" — AI's tendency to serve up crowd-pleasing answers over the hard, operationally necessary choices that real emergencies demand.Timestamps1:37 Go Slow to Go Fast: Minimum governance decisions, the stoplight model, and building your lightweight review circle8:00 Verify Before Trust: Consequence-driven verification workflows, no-go outputs, and two-person integrity for solo shops13:31 Follow the Audit Trail: Where logging lives operationally, keeping it lightweight, and the unintended learning loop16:08 AI Dependency & Human Deskilling: Preserving expert judgment, designing active human-AI workflows, and what to demand from vendors27:27 Trendslop: Why AI favors consensus over context, and how to force trade-offs instead of buzzword-driven incrementalismAboutJonathan Spector is Founder & Principal Advisor at Converge Strategic Partners and a leadership and strategy advisor with more than 20 years of experience helping public-sector, emergency management, homeland security, and cross-sector leaders align priorities, strengthen coordination, and move from strategy to execution. He has worked with government and regional leadership teams on planning, stakeholder alignment, governance, and complex coordination challenges, especially in high-stakes environments where trust, clarity, and execution matter. His current work includes helping leaders navigate AI in practical terms, focusing on governance, guardrails, adoption, and the human side of change.Links:Converge Strategic Partnershttps://www.convergesp.com/Jonathan’s LI profile:https://www.linkedin.com/in/jospector/Leading Through AI for Emergency Management Leaders Executive Reference Guidehttps://www.convergesp.com/ai-executive-reference-guide Research:Harvard Business Review. (2026, March). Researchers asked LLMs for strategic advice. They got “trendslop” in return.https://hbr.org/2026/03/researchers-asked-llms-for-strategic-advice-they-got-trendslop-in-returnLee, E.H., Yin, Y., Jia, N. et al. Relying on AI at work reduces self-efficacy, ownership, and meaning while active collaboration mitigates the effects. Sci Rep (2026). https://doi.org/10.1038/s41598-026-42312-6Snair, J. (2026, March 28). 10 questions (and 7 buzzwords) to ask any AI vendor before you sign. Preppr.ai On the Edge.https://preppr.ai/blog/10-questions-(and-7-buzzwords)-to-ask-any-ai-vendor-before-you-signTank, A. (2026, March 3). Why the next phase of ChatGPT will be specialized. Forbes. https://www.forbes.com/sites/aytekintank/2026/03/03/why-the-next-phase-of-chatgpt-will-be-specialized/



