E34: The Consumer Rights Revolution with Joshua Browder of DoNotPay

Episode Summary

Executive Summary: Josh Browder frames Do Not Pay as an AI-powered consumer-rights platform that fights fees, refunds, and other small injustices ordinary people rarely challenge. He argues AI should be available to consumers if big companies use it, predicts an AI “arms race” that lowers costs and shifts power, and calls for targeted regulation/liability mainly on large platforms and clearly harmful use cases.

Main Topics: Do Not Pay’s evolution from templates to AI (Priority: 5/5): Browder explains how the company began with parking-ticket appeal templates and expanded to over 200 consumer-rights use cases. LLMs replaced brittle rules-based workflows by improving flexibility, success rates, and natural-language interactions. Consumer-rights AI and the “arms race” with companies (Priority: 5/5): The conversation centers on AI bots negotiating refunds, bill reductions, and disputes with companies that are increasingly using AI themselves. Browder sees this as a necessary balancing force that gives consumers the same technological leverage as corporations. Regulation, liability, and disclosure debates (Priority: 4/5): Browder is skeptical of broad AI disclosure rules and licensing for startups, but supports bans or restrictions on clearly harmful uses like debt collection, political impersonation, and AI sentencing. He favors liability on large platforms to force self-regulation. Deflation, automation, and consumer economics (Priority: 4/5): Browder argues AI will reduce service costs across industries, especially customer support-heavy businesses, making goods and services cheaper. He expects substantial job displacement alongside new roles and eventual redistribution pressures such as UBI. Legal, medical, and high-cost sector disruption (Priority: 4/5): The discussion contrasts legal and medical adoption of AI. Browder thinks medicine is more open to augmentation, while law will resist because it is highly regulated and self-protective, especially in criminal and court settings. Platform strategy, security, and technical standards (Priority: 3/5): Browder stresses platform-agnostic product design and lower-friction onboarding, while discussing authentication, scraping, and legal proof requirements. He also endorses open technical standards so consumer and business AIs can interact efficiently. China comparison and consumer expectations (Priority: 3/5): Browder uses China as an example of faster, more seamless consumer service enabled by technology, arguing that Chinese consumers tolerate less friction than Americans and that AI can accelerate service efficiency.

Key Arguments: Consumers should have access to AI if large companies are using AI against them; technological parity is necessary for fairness. Rules-based templates worked for simple disputes but became brittle, detectable, and less effective than LLM-driven natural-language workflows. AI is especially useful for “concentrated benefit, spread out harm” problems where firms profit from small fees imposed on millions of people. Most useful AI regulation should target large platforms and clearly harmful use cases rather than startups or consumers. Disclosure requirements are hard to enforce because AI is already embedded in everyday tools like autocomplete; broad rules may only burden legitimate actors. Liability shifts to big AI platforms would make them self-police harmful behavior more effectively than loose, general regulations. AI will make many services cheaper by reducing customer-service labor and improving efficiency, but savings may be captured first by capital holders unless redistributed. Law will likely resist AI more than medicine because legal work is more zero-sum, highly regulated, and threatened by automation. Consumer adoption and benefits will be uneven: people who embrace AI will gain an advantage, while older or less tech-comfortable users may lag behind. Open technical standards and API-based interoperability could help AIs negotiate with other AIs, reducing friction and making disputes and financial actions more efficient.

Data Points: Company age: Started in 2015; described as almost eight years old in the interview - Browder’s overview of Do Not Pay’s history Consumer-rights use cases: 200+ - Breadth of Do Not Pay products since starting with parking tickets Ticket appeals success: Hundreds of thousands of tickets appealed - Early template-based product traction GPT improvement in bill negotiation: About 10x improvement in success rate from GPT-3 to GPT-4 - Browder says GPT-4 pushes harder in Comcast/utility negotiations OpenAI usage cost for medical price-scraping project: $80,000 - Cost to scrape hospital bills and standardize pricing data for No Surprises Act use case Do Not Pay scale: Team of 7 - Company size while serving hundreds of thousands of subscribers Savings generated: Mid-nine figures annually - Browder’s estimate of current consumer savings from Do Not Pay Desired larger-scale savings: $100 billion/year - Browder’s rough ambition if consumer AI adoption scaled massively Estimated customer-service share of corporate costs: ~10% - Used to argue AI could meaningfully lower prices across the economy Current AI adoption awareness: 12% tried ChatGPT; 54% heard of it - Cited as evidence that adoption is still early and uneven Projected white-collar job loss: ~30% within 5 years - Browder’s near-term labor displacement estimate Projected total job loss: ~60% within 15 years - Browder’s longer-term labor displacement estimate Robocall damages: $1,500 per call - Statutory recovery referenced for robocall litigation use case Price-protection settlement context: Tens of thousands of dollars - Example of a user who sued robocallers full-time and financed home repairs Consumer AI pricing pressure: GPT-3.5 is 10x cheaper than GPT-4 - Used to recommend model routing by task complexity

Pivotal Quotes: "My admission is if the big companies are using AI, consumers should have access to it too." — Josh Browder: Core fairness principle for Do Not Pay’s consumer AI mission "It really is an arms race to keep getting these successes for consumers." — Josh Browder: Describing the escalating bot-versus-bot dynamic in refunds and bill disputes "I think we need like open AI regulations in the same way we need kind of open banking in Europe and even in the US, where you can have AI transact with other AIs and there's a technical standard where they can communicate." — Josh Browder: Proposal for interoperable standards so consumer and company AIs can work together

Implications: Consumer AI is likely to spread fastest in high-friction, low-dollar disputes, forcing companies to automate too. Expect cheaper services, more liability battles, and pressure for rules that protect users without freezing startups.