AI Won't Make Law Cheap Without Structural Change
The promise that artificial intelligence will democratize legal services rests on a hopeful assumption: that better technology automatically translates to better access. Arvind Narayanan and Sayash Kapoor challenge this assumption directly, arguing that AI capability alone cannot overcome the structural bottlenecks embedded in how legal services are delivered, regulated, and contested.
The Three Bottlenecks
Narayanan and Kapoor identify three barriers that stand between AI advancement and meaningful legal reform. First, unauthorized practice of law regulations prohibit nonlawyers from performing legal work, creating liability risks for organizations that deploy AI tools in legal domains. As Narayanan and Kapoor writes, "Without reforms, if consumers cannot access AI capabilities or lawyers are not incentivized to use AI well, AI will not help people accomplish their legal goals, regardless of how advanced it becomes."
Second, the adversarial structure of American litigation means that productivity gains don't necessarily reduce costs. When both sides gain access to better tools, the competitive equilibrium shifts upward rather than downward. Narayanan and Kapoor explains: "In a world with advanced AI, achieving the same result—like settling favorably or prevailing at trial—would require a greater quantity and quality of legal work."
"Even as productivity increases and cost per legal task falls, parties are locked into an arms race of increasing amounts of legal work required to reach the same outcome."
Third, human involvement remains a limiting factor. Judges, lawyers, and clients all operate at human speeds, placing an upper bound on how much AI can accelerate legal processes without sacrificing meaningful participation.
Why Legal Services Are Expensive
The authors trace high legal costs to three structural factors. Legal services are credence goods—their quality is difficult to evaluate even after consumption. The value of legal work is often relative to the opposition's quality, not absolute. And professional regulations limit competition from alternative business models that could leverage economies of scale.
Partner hourly rates at large law firms now exceed $1,300, up 5.1 percent from 2023. Fortune 200 companies reported average litigation costs nearly doubled over eight years, climbing from $6 million per company in 2000 to $15 million in 2008. Narayanan and Kapoor notes that "these regulations require lawyers to serve clients through partnerships fully owned and financed by lawyers," deterring alternative models that could deliver services at $20–$50 per hour instead of $260.
The Debt Collection Example
The debt collection context illustrates the bottleneck clearly. From 1993 to 2013, debt collection lawsuits grew from 1.7 million to about 4 million. More than 70 percent of defendants lose by default for failing to respond, even though many cases are meritless and responding is not complicated.
An AI system could help defendants navigate these suits. But organizations risk violating unauthorized practice of law laws whenever their AI tools complete tasks requiring legal judgment. Narayanan and Kapoor writes: "While AI's legality remains in doubt, the threat of UPL liability can inhibit its adoption."
LegalZoom's history demonstrates how these regulations deter innovation. The company has faced lawsuits in Missouri, North Carolina, California, and New Jersey over allegations that its automated document preparation constitutes unauthorized practice of law.
Critics Might Note
Critics might argue that the authors underestimate how quickly regulatory frameworks can adapt to technological change. State bar associations have begun issuing guidance on AI use, and some jurisdictions are experimenting with limited licenses for legal technicians. The bottleneck may be temporary rather than permanent.
Critics might also note that not all legal work is adversarial. Estate planning, routine contract review, and compliance work don't involve opposing counsel. AI could deliver significant cost reductions in these domains even without structural reform.
Institutional Reforms Required
The authors conclude that AI will deliver better outcomes only if the legal industry enacts reforms addressing the bottlenecks. Otherwise, "we risk a future in which legal work becomes more abundant, but legal outcomes remain expensive and inaccessible."
Narayanan and Kapoor applies the "AI as Normal Technology" framework to the legal domain, emphasizing agency over predetermined capability trajectories. Better models have not yet translated into more reliable legal products because adapting workflows and teaching users takes time.
Bottom Line
AI capability advances alone will not make legal services cheaper. The adversarial structure of litigation, regulatory barriers to nonlawyer service provision, and the necessity of human involvement create bottlenecks that technology cannot overcome without institutional reform. The legal profession faces a choice: enact structural changes that allow AI to improve access and efficiency, or accept a future where legal work becomes more abundant but legal outcomes remain inaccessible to those priced out of the system.