NBER Trial: AI Boosts Junior Patent Attorneys Now, Erodes Skills Later

A three-month randomized controlled trial with 133 patent attorneys across eleven U.S. intellectual property law firms found that AI assistance improves immediate performance most for junior lawyers — but only senior attorneys retain any benefit once the AI is taken away.

The working paper comes from the National Bureau of Economic Research (NBER), a nonprofit U.S. economics research organization, and was funded by Google. It addresses a contested question in current research: whether AI assistance builds or erodes professional expertise.

The stakes extend well beyond patent law. Firms adopting AI tools gain productivity immediately, while the cost of stunted talent development may spread across entire professions — a dynamic some researchers call the "tragedy of the cognitive commons."

How the experiment worked

Two-thirds of the participants received access to "InFlow," an unreleased AI assistant from Google Labs. The remaining third worked without any AI as a control group.

The researchers separated two measurements. Attorneys first worked with AI assistance, then evaluated results without it. This design let the team measure both the quality of AI-assisted work and how well each attorney's own judgment held up afterward.

The tasks broke down as follows:

• In two patent claim drafting tasks, only the treatment group could use InFlow; the control group completed the same tasks unaided.
• In a third task called redlining, all participants had to critically review an existing patent draft without AI.

As the authors explain, drafting measures immediate AI-assisted performance, while redlining measures internalized judgment — a task patent attorneys routinely perform without any tools.

Blinded patent attorneys at an independent firm scored all results across five categories: enforceability, technical accuracy, strategic ambiguity, completeness, and clarity.

Where did AI help most?

In the drafting tasks, the treatment group clearly outperformed the control group at both the 10-day and 90-day marks, with effects showing up across all five scoring dimensions. Higher scores came mostly from fewer bad ratings and more mid-range ones, while top scores barely moved.

Gains were much larger for juniors with fewer than seven years of experience than for seniors. According to the researchers, this aligns with a growing body of studies showing that AI narrows performance gaps between less and more experienced workers — an effect already observed in writing, customer service, and software development.

What happened when the AI was removed?

Redlining produced the opposite result. The treatment group scored better overall, but the advantage came entirely from senior attorneys. Juniors saw no benefit at all, even though they had posted the biggest gains during AI-assisted drafting.

Junior results split apart. Far fewer mid-range scores appeared, replaced by more bad and more good ones, with no growth at the top. The authors read this as a sign that AI sped up skill-building in some juniors while replacing it in others, though the experiment does not explain why the effects diverged.

The pattern carries a sharp implication: the attorneys who gained the most from AI were the same ones who retained the least lasting skill. The researchers suggest baseline expertise may be a prerequisite for turning AI-assisted practice into durable ability.

A supplementary text analysis offers one possible explanation for the gap. Senior attorneys more often restructured the flawed draft and grounded their changes in legal principles. Juniors frequently spotted serious flaws without fixing them, leaving notes that read like instructions to an AI assistant. Because this behavior appeared in the control group too, the authors do not attribute it to an AI-trained habit. They see it as a baseline deficit at the junior level that AI use did not fix.

What are the study's limitations?

The authors point to several constraints on their findings:

• A small sample size, driven by top attorneys' high billable rates.
• Participating firms are not representative of professional services as a whole.
• The study lacked a baseline skills measurement.
• The three-month window is short compared to the years it takes to build patent expertise.

Why the tension matters for professional work

The researchers argue that the tension between immediate productivity gains and lasting skill development matters as long as professional work still requires judgment without AI — and it likely will for a long time. Even if AI eventually supports redlining, it probably will not be available to attorneys in real time during client meetings and courtroom proceedings, where stakes are high and judgment calls happen on the spot.

Similar patterns already show up in other fields:

• In software development, an Anthropic study found that younger developers who learned a new programming library with AI help scored 17 percent worse on a follow-up knowledge test than those who learned without it.
• In essay writing, the MIT Media Lab found signs of cognitive debt from outsourced thinking.
• In management consulting, a BCG study showed back in 2023 that weaker consultants improved far more with GPT-4 than strong performers did.

Stanford researchers explain the decline in entry-level jobs across multiple industries with a related argument: AI can mostly replace textbook knowledge from school and college, which is the main thing early-career workers bring to the table. Experienced professionals also carry hands-on expertise that AI cannot easily substitute.

The labor-market signal is already visible. Anthropic recently said that junior positions in its engineering departments dropped from 21 percent to 14 percent since early 2025.

The NBER paper adds experimental evidence from real-world legal practice to the broader debate on the erosion of professional expertise. Its core finding sets the terms for that debate in concrete terms: firms that hand juniors an AI assistant collect the drafting gains now, while the profession absorbs the deferred cost of attorneys who never built the judgment the tool was supposed to augment.