National Bureau of Economic Research working paper<\/a>, which has not been peer-reviewed. According to the paper, it enrolled 133 lawyers at 11 U.S. intellectual-property law firms that have ongoing patent-drafting relationships with Google. Researchers randomly assigned access to a custom AI patent-drafting assistant, a then-unreleased Google Labs tool. Two-thirds of the lawyers got access. The rest went without until the study ended.<\/p>\nGoogle\u2019s role went beyond the tool. The paper lists Autor\u2019s six co-authors as Google employees. It also states that Google conducted the study and paid its direct costs, and that MIT\u2019s human-subjects committee determined MIT was not engaged in the research. In an email to Fortune, Google representatives described the paper as independent research Autor did as part of his fellowship.<\/p>\n
Lawyers drafted patents from simulated inventor materials after 10 days and again after 90 days. Patent attorneys at an independent law firm graded the work without knowing which lawyers had used AI, according to the paper. They scored enforceability, accuracy, strategic ambiguity, completeness and clarity.<\/p>\n
After 90 days, AI access raised drafting scores by 0.38 standard deviations, the paper found. In a Google blog post accompanying the paper, Autor and co-author Tanya Rodchenko described that as an 11-percentile-point gain relative to control scores. The paper says the improvement came from less weak work, not more excellent work.<\/p>\n
The harder test came at the end. Lawyers had to mark up a hypothetical patent that, according to the blog post, contained many substantive and stylistic errors. The paper calls this a task patent lawyers \u201croutinely perform unaided.\u201d No AI was allowed.<\/p>\n
Senior lawyers who had AI access beat their control-group peers by 0.45 standard deviations, according to the paper. Junior lawyers showed no average gain.<\/p>\n
The paper also reports caveats. Only 91 of the 133 lawyers finished the AI-free test. The firms declined a skills test at the outset, so the researchers inferred skill gains from random assignment instead of measuring before and after. The authors also flagged 15 of the 91 markups as possibly AI-assisted despite the ban, and the flags turned up about as often among control lawyers. With those 15 excluded, the paper says, the senior advantage shrinks to 0.39 standard deviations and holds up, though at a looser statistical threshold. The overall effect does not.<\/p>\n
The \u2018illusion of competence\u2018<\/p>\n
The juniors\u2019 scores split, the paper found: significantly more poor scores, fewer mediocre ones, more good ones and no gain at the top. The authors write that AI \u201cserved as a springboard for some juniors and a cushion for others.\u201d<\/p>\n
The blog post describes how the juniors worked. They tended to go top to bottom, polishing introductory text before reaching the main claims. Some spotted serious flaws but left comments describing them instead of fixing them. Control-group juniors did the same, a pattern the authors call \u201ca baseline junior deficit\u201d that three months of AI access did not fix.<\/p>\n
The juniors did notice a difference: they liked the tool. In his responses to Fortune, Autor said AI-assisted junior lawyers reported a surge in task satisfaction, the largest effect in the study. They felt that skipping the \u201cblank page problem\u201d let them step straight into a reviewer role, he said.<\/p>\n
\u201cYoung professionals need to beware of the illusion of competence,\u201d he said. \u201cThe only way you\u2019re really going to know if you\u2019re developing skills is if you do tasks without AI assistance and evaluate your performance.\u201d<\/p>\n
Why the older lawyers gained<\/p>\n
According to the blog post, the experienced lawyers described AI in follow-up interviews as a \u201clogic auditor\u201d and not as a finished product. The authors write that it weakened their attachment to existing prose and forced them to spell out the why and how of their structural edits.<\/p>\n
\u201cArguably, you can use AI as a \u2018logic auditor\u2019 only if you already know the law well enough to spot when fluent text is legally flawed,\u201d Autor told Fortune. He said he believes offloading baseline drafting freed seniors to focus on strategic scope, \u201cthe forest not the trees.\u201d<\/p>\n
On the unassisted test, the blog post says, treated seniors skipped low-stakes prose, rebuilt claims from scratch, cut language that could narrow legal protection and tied many of their edits to legal doctrine.<\/p>\n
That doesn\u2019t mean juniors should abandon the tool. Autor told Fortune that novices should work like seniors: frame the argument unassisted first, then use AI as a critic to pressure-test the logic. He acknowledged that is a lot to ask. \u201cMost of us don\u2019t have the self-discipline to do things by hand when a readily available, low-effort tool is sitting right there, offering to do it for us,\u201d he said.<\/p>\n
Don\u2019t cut the apprenticeship<\/p>\n
For employers, better output creates a temptation to hire fewer juniors. Autor called that short-sighted. \u201cIf firms automate away formative practice without replacing it with some kind of guided learning environment, they sever the apprenticeship pipeline that produces tomorrow\u2019s senior partners,\u201d he told Fortune.<\/p>\n
He suggested firms decide which tasks can be permanently completed with AI assistance and which judgment tasks require unassisted mastery. They could also pair AI rollouts with regular unassisted skill checks, such as offline redlining exercises, and hold partners accountable for mentoring juniors through the technology\u2019s mistakes. These are ideas, not proven fixes. \u201cIt\u2019s early days in this area of research and practice,\u201d Autor said, and few solutions have been robustly proven.<\/p>\n
The paper lists other limits. The sample was small and drawn from firms doing ongoing patent work for a single sophisticated client, and three months is brief compared with the years it takes to build patent expertise. The blog post adds that newer models and wider AI familiarity might change the results if the trial were run today. It does not show that AI keeps juniors from becoming experts over a career.<\/p>\n
Autor ties the issue back to his China work. Most technological advantage, he told Fortune, comes less from equipment or blueprints than from human know-how, built through slow, laborious mastery. \u201cIf we think that AI will relieve us of the burden of mastering expertise, I think we will be sorely disappointed,\u201d he said. \u201cHuman intelligence and machine intelligence will be complements for the long term.\u201d<\/p>\n","protected":false},"excerpt":{"rendered":"Gen Z was promised that AI would be the great equalizer: a tool that lets a 24-year-old produce…\n","protected":false},"author":2,"featured_media":798526,"comment_status":"","ping_status":"","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[20],"tags":[554,733,4308,20635,70158,844,1215,86,56,54,55],"class_list":["post-798525","post","type-post","status-publish","format-standard","has-post-thumbnail","category-artificial-intelligence","tag-ai","tag-artificial-intelligence","tag-artificialintelligence","tag-automation","tag-disruption","tag-google","tag-law","tag-technology","tag-uk","tag-united-kingdom","tag-unitedkingdom"],"_links":{"self":[{"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/posts\/798525","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/comments?post=798525"}],"version-history":[{"count":0,"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/posts\/798525\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/media\/798526"}],"wp:attachment":[{"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/media?parent=798525"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/categories?post=798525"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/tags?post=798525"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}