From Red Tape to Green Tape: Decluttering the State with AI and Collective Intelligence

In 2011, while working with 10 Downing Street on government modernization, I helped launch the Green Tape Challenge. Four years later, in Texas, I designed the Red Tape Challenge for the House Government Efficiency and Reform Committee. In both cases, the premise was simple: invite the public to identify outdated rules and regulations and propose modern alternatives that would achieve the same public purpose with less burden.

In both cases, the premise was simple: invite the public to identify outdated rules and regulations and propose modern alternatives that would achieve the same public purpose with less burden.

The need was obvious. Laws and regulations accrete over decades — centuries, in the case of the United Kingdom — and too often persist long after the conditions that gave rise to them have changed. In Texas, we focused on four areas: occupational licensing, state agency rulemaking, public school mandates, and manufacturing. We asked residents to offer concrete recommendations, from high school curriculum mandates to regulations governing public insurance adjusters. As the Speaker of the House put it at the time, “The direct input provided on this site will help the Legislature make state government more efficient, accountable, and responsive to taxpayers.”

Laws and regulations accrete over decades — centuries, in the case of the United Kingdom — and too often persist long after the conditions that gave rise to them have changed.

Such efforts, however, were the exception rather than the rule and they relied exclusively on crowdsourced input without the benefit of AI. 

Today, governments remain mired in what Jennifer Pahlka has aptly called “policy clutter.” From Virginia to Ohio to San Francisco, jurisdictions are now turning to artificial intelligence to turbocharge regulatory simplification and streamlining. AI systems can scan vast regulatory corpora, identify duplication and conflict, and surface opportunities for modernization at a speed and scale previously unimaginable. But these new efforts are missing the benefits that stemmed from public participation.

That is why regulatory reform must combine artificial intelligence with collective intelligence. AI can perform the first pass — mapping complexity, identifying friction, surfacing patterns invisible to human review. But people are needed to supply context, values, and judgment: to explain lived experience, to articulate public purpose, and to co-design solutions that execute regulatory intent better rather than abandon it altogether.

Ohio’s recent experience offers a glimpse of the promise of AI-driven reform. The state's Common Sense Initiative deployed AI tools to analyze millions of words of statutory and regulatory text, identifying duplicative higher ed regs, trimming hundreds of thousands of words from the building code, and more than a million words of obsolete lottery regulations, with projected savings of tens of millions of dollars and tens of thousands of staff hours. Yet the process largely stopped at identification and elimination, leaving unanswered questions about whose burdens were reduced, which protections were preserved, and who had a voice in deciding.

As I learned firsthand in both the UK and Texas, these decluttering initiatives carry risks. Calls to “simplify” regulation can mask political efforts to gut consumer protections, weaken environmental safeguards, or shirk core government responsibilities. Efficiency, untethered from legitimacy, becomes a pretext for eroding the public interest rather than strengthening it.

Calls to “simplify” regulation can mask political efforts to gut consumer protections, weaken environmental safeguards, or shirk core government responsibilities. Efficiency, untethered from legitimacy, becomes a pretext for eroding the public interest rather than strengthening it.

These results matter. But they also point to a deeper truth: AI alone cannot tell us which burdens matter most, to whom, or at what cost. Algorithms can flag redundancy and contradiction. They cannot explain how delays cascade for a small business, how discretionary enforcement feels arbitrary to a resident, or how compliance costs disproportionately affect those least able to bear them. Nor can they decide which safeguards are essential and which are vestigial.

As we look for new ideas for how to declutter and improve the regulatory state, we have to look to solutions that combine artificial intelligence with collective intelligence. AI can perform the first pass — mapping complexity, identifying friction, surfacing patterns invisible to human review. But people are needed to supply context, values, and judgment: to explain lived experience, to articulate public purpose, and to co-design solutions that execute regulatory intent better rather than abandon it altogether.

In October, Political Watch and MySociety convened the Designing the Green Tape Challenge workshop to explore what this combination could look like in practice. Participants worked through how AI tools might help identify regulatory burden, while collective deliberation ensures that reforms protect health, safety, equity, and environmental goals.

Together, the group set out to answer a deceptively hard question: How might the City of Boston replace red tape with green tape — governance that is simpler, fairer, and smarter without sacrificing the public interest? What they produced over the course of the brief but intense "idea-a-thon" was not a single technical solution, but a structured process, designed to ground modernization in purpose rather than mere efficiency.

Several themes emerged in what the group suggested for Boston:

First, participants emphasized that AI is best suited to diagnosing “high-effort” rules, not judging “low public value.” Models can reliably detect complexity, duplication, excessive steps, unclear language, long wait times, and procedural bottlenecks. Determining whether those burdens are worth bearing, however, requires human judgment grounded in lived experience and public values. As several participants noted, public value is not a technical attribute: it is a democratic one.

Second, the group stressed the importance of distinguishing local discretion from higher-level legal constraints. Municipal permitting systems are often shaped by state, federal, or even supranational rules that cities cannot change. Any AI-enabled reform effort must therefore map not only procedural burden, but also jurisdictional authority, so that public participation is directed toward decisions that are actually within local control rather than becoming an exercise in frustration.

Third, participants challenged the assumption that simplification should always proceed by shared purpose alone. In some cases, integrating permits with similar goals makes sense. In others, the more meaningful simplification comes from aligning processes around shared users or workflows  (e.g. consolidating touchpoints for residents or businesses even when underlying purposes differ). This insight underscored how AI-generated clustering must remain provisional, subject to human reinterpretation.

Fourth, the group devoted significant attention to how participation itself must be redesigned. Participants warned against symbolic consultation and argued instead for structured co-design: combining AI-enabled synthesis with deliberative forums, workshops, and in-person engagement. They highlighted the need to reach those most burdened by permitting systems — immigrants, small business owners, low-income residents — by going to where they already are, including permitting offices themselves, and by offering meaningful recognition or incentives for participation.

Finally, the workshop reinforced that “green tape” is not deregulation. The goal is not fewer rules, but rules that are clearer, fairer, and better aligned with their original intent. Participants consistently returned to the need for human oversight, transparency in AI reasoning, mechanisms for appeal and correction, and safeguards against bias in both data and participation.

Taken together, the Green Tape Challenge points toward a model of regulatory reform that treats AI as an enabler of democratic problem-solving rather than a substitute for it. In a next installment, I'll write more about what Boston has gone on to build (spoiler alert: it combines CI and AI). 

By pairing machine analysis with collective judgment, governments can reduce unnecessary burden while strengthening — not weakening — the public interest.

A fuller account of the workshop submissions, design proposals, and open questions is available here.