Episode Overview
Podcast: The AI Policy Podcast
Host: Gregory C. Allen, Center for Strategic and International Studies (CSIS), Wadhwani Center for AI and Advanced Technologies
Guest: Jennifer Pahlka (Former US Deputy CTO, Founder of Code for America, Founder of Recoding America Fund, Author of "Recoding America")
Title: Jennifer Pahlka on Reforming Government for the AI Era
Date: February 5, 2026
Main Theme:
This episode explores the intersection of government reform, technology, and artificial intelligence (AI), focusing on how bureaucratic and policy processes impede progress and how AI can both exacerbate and help tackle these challenges. Jennifer Pahlka shares insights from her career, the origins and evolution of US government digital initiatives, and offers deep reflections on how to truly unlock government’s capacity to serve citizens in the AI era.
Key Topics & Insights
1. Jennifer Pahlka’s Journey to Government Reform
-
Early Experience with Government Inefficiency
- Pahlka’s first job (1992) at a child welfare agency exposed her to inefficient, paperwork-heavy processes.
- Experience in tech (Game Developers Conference, Web 2.0 Conference) informed her later push to modernize government digital services.
-
The Birth of “Gov 2.0” and Digital Service Initiatives
- Collaboration with Tim O’Reilly and others defined a movement focused on lightweight, user-centered tech practices for government (Gov 2.0).
- Founding of Code for America to bring these principles into state and local government.
2. The U.S. Digital Service (USDS) & The Healthcare.gov Crisis
- Genesis of USDS
- Inspired by UK’s Government Digital Service (GDS) but tailored for US government.
- Healthcare.gov’s catastrophic launch was the catalyst: “Retroactively we … called that the first project of USDS… Are they there for firefighting or are they there to change practices more sustainably in the long-term?” (A, 09:37–10:50)
3. From “Tech Will Fix It” to “Policy and Process Must Change”
-
Recognizing Structural Barriers
- Inserting talented technologists is necessary but insufficient.
- “You have to go upstream to the structural reasons that we are getting bad outcomes… It is still the operating model from the post-World War II industrial era. We have sort of slapped websites on the front end of that and pretended that it’s a fit for purpose for the Internet era. But we didn’t do the backend work to really update it.” (A, 12:27–13:32)
-
Upstream Bottlenecks Example (DoD Employee Onboarding)
- “We’ve got this incredible Ferrari of a brain and he’s basically forbidden from using his skills in a meaningful way… because of the cybersecurity paradigm.” (B, 14:05–15:02)
4. Cultural and Practical Obstacles in Government
-
Rigidity & Over-Process
- Government culture is strongly rule-following, often for legitimate reasons (e.g., nuclear safety), but process can become divorced from outcomes.
- “We just know how to add, we don't know how to subtract. And that is the skill that we need to learn.” (A, 52:02–53:02)
-
The “Cascade of Rigidity”
- Over time, original legislative intent becomes lost under layers of regulation, memos, and risk aversion.
- Memorable Quote: “I call this thing the cascade of rigidity. As it’s gone from what lawmakers at a very high level sort of stated their intent, as it's gone down through the hierarchy and been implemented, it's become more and more rigid at every step and gotten sort of locked into this practice ...” (A, 26:04–27:41)
-
Example - VA/Irs and Fax Machines
- Resistance to change based on presumed legal requirements—often misinterpretations or outdated rules.
- AI can help determine whether a rule is genuinely required by law or simply accreted bureaucracy.
5. The Potential of AI for Government Reform
-
AI as a Tool for Policy Analysis & Process Simplification
- “AI is an incredible tool to help us determine: is this thing … actually deriving from law that we need Congress to go change? Or can somebody … change this without any fear that we have actually broke [the law]?” (A, 28:37–29:09)
- Use cases include searching and analyzing vast regulatory texts, detecting redundant or conflicting rules, and aiding in large-scale policy simplification (e.g., New Jersey’s unemployment insurance rules).
-
Beyond Tech: Asking the Right Questions
- AI should not just optimize or automate existing, possibly outdated functions, but prompt reconsideration: "Are these the right programs to be delivering?" (A, 37:55–38:45)
- Government must adopt test-and-learn frameworks—constant iteration, not “set-it-and-forget-it.”
Notable AI Use Cases Highlighted:
- Automating Legal and Regulatory Review: Fast, accurate searches through complex law to spot opportunities for reform.
- Workforce and Policy Modernization: Identifying vestigial job classifications or rules (e.g., Maryland’s 4,500 job categories).
- Assisting Human Judgment in Adjudication: Custom models to streamline complex, human-judgment-heavy processes like unemployment insurance.
6. The Four Pillars for Modernizing Government (A, 41:01–50:07)
-
Recruit and Hire the Right People
- “It was Workforce all along. Every time it’s the HR systems, like any persistent problem that you have... it’s HR problems.” (A, 43:01–43:48)
-
Focus on the Right Work
- Procedural reform to eliminate unnecessary burdens and focus on outcomes (“the comically misnamed Paperwork Reduction Act”).
-
Build Purpose-Fit Systems
- Reform how we build and buy technology; focus on fixing processes, not simply increasing resources (“unkink the hose”).
-
Test and Learn Frameworks
- Continuous iteration, learning from results, adapting processes and technology accordingly: “If you're using AI, good Lord, please don't ever let your testing be over. You should be constantly monitoring and improving these systems.” (A, 50:00–50:07)
7. Tech for Tech’s Sake vs. Outcomes
-
Avoiding "Technology Fetishization"
- “You care about outcomes, you care about effectiveness. … Whether old or new or weird or common … what is going to deliver these outcomes?” (B, 53:12–54:23)
-
Emphasis on User Needs and Trust in Government
- “People expect government to work well… and so we need to meet. We also just need to, like, make them feel like government is trying and using their tax dollars well, so that people have trust and faith…” (A, 55:25–57:44)
Notable Quotes & Memorable Moments
-
On Process vs. Outcomes
"There's a culture of rule-following that exists for a good reason. But this sort of shift from the religion about the process to focusing on the outcomes is so desperately needed..."
— Gregory Allen, [22:09] -
On AI for Regulatory Navigation
“I’m very excited about AI’s ability to just empower every public servant to ask that question [does the law require this?] so that they can... figure out the answer.”
— Jennifer Pahlka, [28:56] -
On Legacy Constraints
“We just know how to add, we don’t know how to subtract. And that is the skill that we need to learn to start making these systems much more stable and scalable.”
— Jennifer Pahlka, [53:02] -
On Government Innovation
“You want to attract the best and the brightest guy in... like they don’t have it. Like the needs have changed.”
— Gregory Allen, [41:50] -
On Bipartisanship
“Both parties have been a source of [these problems]. And I think both parties are now very curious about how to solve them in slightly different ways, but in ways that have enough overlap that it makes me very optimistic.”
— Jennifer Pahlka, [45:38–46:23]
Timestamps: Important Segments
- [04:03] – Jennifer Pahlka’s early career and entry into government & technology
- [07:13] – USDS origins, healthcare.gov crisis as a turning point
- [11:42] – Upstream process & policy blockers, shift away from “add more tech”
- [15:28] – Practical obstacles: hiring delays, IT access, and hardware issues
- [18:00–20:37] – Budgeting issues and systemic inefficiency (CAC bottlenecks)
- [24:13–26:04] – Regulatory rigidity and the legal basis for rules
- [30:49] – AI as a tool for policy analysis and simplification
- [33:52–37:12] – Opportunities for efficiency and improved outcomes via AI
- [41:01–50:07] – The Four Pillars of Government Reform
- [53:02] – The necessity of learning to “subtract,” not just add, in policymaking
- [57:44] – The narrative on trust, user needs, and outcomes in government
- [59:29–64:35] – AI applications for process reform; job classification example
- [65:48–68:25] – Recommendations for further reading & resources
Recommended Reading & Resources (Curated by Jennifer Pahlka) [65:48]
- Recoding America (her own book)
- Hack Your Bureaucracy – Marina Nitze & Nick Sinai
- Crisis Engineering (forthcoming) – Marina Nitze and colleagues
- The Procedure Fetish – Nick Bagley (Niskanen Center)
- Why Nothing Works – Mark Dunkelman
- Kill It With Fire – Marianne Bellotti
- Phoenix Project and The Goal – novels structured as business optimization case studies
Additional Resources
- Jennifer Pahlka’s Substack: Eating Policy
- Recoding America Fund: New initiative aiming to build tools for policy analysis and reform
Tone & Style
Throughout the episode, the tone is conversational, insightful, pragmatic, and optimistic. Pahlka communicates with humility, practical expertise, and a resolute focus on solving real systemic issues, while Allen brings out policy nuances and shares personal “in the trenches” stories that highlight the urgency and complexity of government digital reform.
SUMMARY
This episode is an essential listen for anyone interested in AI policy, government modernization, or organizational reform. Jennifer Pahlka’s deep experience and clear-headed approach reveal how government’s biggest barriers are not always technical, but deeply rooted in culture and process. AI has a transformative role to play—not in “automating for automation’s sake,” but in unlocking government’s true capacity to deliver for citizens, provided reformers keep their eyes on outcomes and the human element at the heart of public service.
