One of the recurring blind spots in public sector AI enthusiasm is a failure to answer a basic question: Why would governments succeed with GenAI now, given their long history of struggling to adopt much simpler technologies?
In other words, few advocates can articulate their implicit leapfrogging hypothesis. (As a general rule, I’ve found that the more enthusiastic someone is about GenAI in the public sector, the less able they are to answer that question.)
My view is that any serious attempt to harness GenAI in the public sector must start with two questions:
- Why did previous tech waves fail?
- What new affordances make this different?
Dave Guarino’s post offers one compelling example of how GenAI may shift the terrain: for the first time, frontline staff can evaluate tools on their actual problems for a few dollars in API fees, potentially breaking the old “procure and pray” cycle, which often leads to waterfall projects that inevitably fail.
But this is just one potential advantage. I’m curious what other structural differences people see that might enable leapfrogging this time.
What makes GenAI different from previous waves of government tech adoption? Are there other affordances technical, economic, or organizational that address the root causes of past failures?
I’m particularly interested in concrete examples like Dave’s, where you can point to a specific barrier that historically killed projects and explain how GenAI’s characteristics might circumvent it.
What am I missing? What other reasons might make this wave different?