Most revolutions in legal technology rely on the next generation to come in and save the seasoned Luddites from themselves. From desktop computers to the internet, it’s the incoming lawyers who performed the heavy lifting while the senior partners continued printing out emails and sending handwritten replies through interoffice mail.
But as the legal industry dreams of its AI-native future, the next generation of lawyers are a lot less psyched about the technology’s potential. While firms allocate millions (if not a billion) to developing AI tools to take work from the first touch to fully drafted work product, college students don’t trust AI much beyond a fancier research tool.
“The College Student AI Reality Check,” is a new report from Blickstein Group reflecting responses from students across 37 colleges and universities, plus five in-depth interviews. Spearheading the report was Vera Grabow, the group’s intern and a college junior double majoring in History and Political Science, who launched the project based on her own experience watching AI reshape her education and recognizing that no one had actually asked students what they thought about the technology that every press release breathlessly proclaims to be the future of work. “She conceived this project, convinced us to undertake it, wrote the questionnaire, lead the outreach, analyzed the data, and drafted the report,” CEO Brad Blickstein said. “We helped, but this was an exceptional piece of work by a college student.”
The good news for the industry is that students are using AI. Some 71 percent of respondents reported using AI several times a week or more for their schoolwork. Before professors clutch their pearls too tight, the students aren’t turning their assignments over to ChatGPT:
In fact, AI has replaced Google as the preferred first stop for figuring out a new concept, with 27.8 percent reporting that they ping an AI tool first, edging out Google at 25.3 percent. Not that Google isn’t just an AI answer followed by six sponsored links at this point. Both destinations far outpace going to professors and TAs, which clocked in at 3.8 percent. For faculty, office hours remain the best place on campus to be alone with your thoughts.
Research, proofreading, flashcards, brainstorming… those aren’t use cases that excite the greater AI industry. The investors pouring money into the frontier lab furnace expect a technology that will fully replace human labor and allow them to realize their dream of turning poor people into Soylent Green. Meanwhile, the young people most engaged with technology are wary of trusting it as much more than a better search bar and souped up grammar checker.
“I think it’s improved my performance,” a Stanford Earth Systems senior cited in the report said. “I don’t think it has improved my own knowledge.”
Few doubt the technology as a productivity enhancer, but that’s not exactly in line with the “agent drafts the motion while you sleep” pitch that the OpenAI needs if it ever hopes to see enough tokens burned to make ends meet.
The key metric for legal tech is that “acceptable for drafting if substantial edits are made” and less than 8 percent is a brutal number. Employers in every industry will do their best to train its new people to trust the bots, but habits and preconceptions die hard. If most students come out of college skeptical of the tech, it’s going to be hard to either recruit AI enthusiasts or easily rewire disbelievers.
When asked how much time AI actually saves them on a typical assignment, the most common answer is 10-24% of time saved (32.4%), followed by 25-49% (28.2%). Only 8.5% report saving 75% or more, and an identical 8.5% report saving less than 10%. For a tool that many students use on a regular basis, the payoff they describe seems more like an assistant and aid that creates incremental improvements than transformative solutions.
And so much for the technology making itself indispensable in the minds of users:
The report notes that not a single respondent chose “not at all confident.”
A whole 72.1 percent believe heavy reliance on AI leaves people “somewhat” or “much” less capable at their jobs.
Not that students aren’t occasionally doing what they know is probably a bad decision. One respondent had AI digest a two-hour documentary for class rather than watch it. One of the above graphs noted that 46.5 percent use AI to summarize readings. It’s one thing to appreciate a summarizing tool, and another to farm out reading comprehension.
However grim these numbers are for AI, they’re even worse when narrowed to the sample of pre-law students:
These students use AI as often as other respondents, yet are much more skeptical of it: 66.7% say AI is acceptable for brainstorming only, against 38% of the full sample; 66.7% believe heavy reliance will leave them much less capable, against 31.6% overall. Their confidence that they could work without AI is higher (83.3% versus 62%), and their appetite for a collective AI-free semester is 100%, against 80.3% for everyone else.
“A group that uses AI often and trusts it least is not a contradiction,” the report notes, putting these figures in perspective. “Rather, perhaps it mirrors the profession’s instincts showing up early in students who appreciate traits that the law values.” That’s good news for the profession, where AI’s most responsible use case — putting aside true brute force tasks like doc review — is as a teammate overseen by a skeptical human. But it’s not great news for the frontier labs underwriting the supposed revolution.
If AI evangelists want to usher in a new era of work, they’re going to need to spend time persuading this population to trust the technology more. Right now, they’ve got some deeply unsure users who think it’s a fine tutor, but a threat to their own competence. Unless this group learns to get comfortable with a workflow that includes allowing AI to put together a first draft, it’s hard to imagine them entering the workforce capable of getting the most out of these tools even if they wanted to.
My personal experience talking to college students largely tracks this report. A lot of use… a lot of distrust. But I’d been running on vibes. Grabow lives in that world and decided it’s long past time to gather some data.
Check out the full report here.
Joe Patrice is a senior editor at Above the Law and co-host of Thinking Like A Lawyer. Feel free to email any tips, questions, or comments. Follow him on Twitter or Bluesky if you’re interested in law, politics, and a healthy dose of college sports news.
The post The Next Generation Of Law Students Don’t Trust AI Hype appeared first on Above the Law.
Most revolutions in legal technology rely on the next generation to come in and save the seasoned Luddites from themselves. From desktop computers to the internet, it’s the incoming lawyers who performed the heavy lifting while the senior partners continued printing out emails and sending handwritten replies through interoffice mail.
But as the legal industry dreams of its AI-native future, the next generation of lawyers are a lot less psyched about the technology’s potential. While firms allocate millions (if not a billion) to developing AI tools to take work from the first touch to fully drafted work product, college students don’t trust AI much beyond a fancier research tool.
“The College Student AI Reality Check,” is a new report from Blickstein Group reflecting responses from students across 37 colleges and universities, plus five in-depth interviews. Spearheading the report was Vera Grabow, the group’s intern and a college junior double majoring in History and Political Science, who launched the project based on her own experience watching AI reshape her education and recognizing that no one had actually asked students what they thought about the technology that every press release breathlessly proclaims to be the future of work. “She conceived this project, convinced us to undertake it, wrote the questionnaire, lead the outreach, analyzed the data, and drafted the report,” CEO Brad Blickstein said. “We helped, but this was an exceptional piece of work by a college student.”
The good news for the industry is that students are using AI. Some 71 percent of respondents reported using AI several times a week or more for their schoolwork. Before professors clutch their pearls too tight, the students aren’t turning their assignments over to ChatGPT:
In fact, AI has replaced Google as the preferred first stop for figuring out a new concept, with 27.8 percent reporting that they ping an AI tool first, edging out Google at 25.3 percent. Not that Google isn’t just an AI answer followed by six sponsored links at this point. Both destinations far outpace going to professors and TAs, which clocked in at 3.8 percent. For faculty, office hours remain the best place on campus to be alone with your thoughts.
Research, proofreading, flashcards, brainstorming… those aren’t use cases that excite the greater AI industry. The investors pouring money into the frontier lab furnace expect a technology that will fully replace human labor and allow them to realize their dream of turning poor people into Soylent Green. Meanwhile, the young people most engaged with technology are wary of trusting it as much more than a better search bar and souped up grammar checker.
“I think it’s improved my performance,” a Stanford Earth Systems senior cited in the report said. “I don’t think it has improved my own knowledge.”
Few doubt the technology as a productivity enhancer, but that’s not exactly in line with the “agent drafts the motion while you sleep” pitch that the OpenAI needs if it ever hopes to see enough tokens burned to make ends meet.
The key metric for legal tech is that “acceptable for drafting if substantial edits are made” and less than 8 percent is a brutal number. Employers in every industry will do their best to train its new people to trust the bots, but habits and preconceptions die hard. If most students come out of college skeptical of the tech, it’s going to be hard to either recruit AI enthusiasts or easily rewire disbelievers.
When asked how much time AI actually saves them on a typical assignment, the most common answer is 10-24% of time saved (32.4%), followed by 25-49% (28.2%). Only 8.5% report saving 75% or more, and an identical 8.5% report saving less than 10%. For a tool that many students use on a regular basis, the payoff they describe seems more like an assistant and aid that creates incremental improvements than transformative solutions.
And so much for the technology making itself indispensable in the minds of users:
The report notes that not a single respondent chose “not at all confident.”
A whole 72.1 percent believe heavy reliance on AI leaves people “somewhat” or “much” less capable at their jobs.
Not that students aren’t occasionally doing what they know is probably a bad decision. One respondent had AI digest a two-hour documentary for class rather than watch it. One of the above graphs noted that 46.5 percent use AI to summarize readings. It’s one thing to appreciate a summarizing tool, and another to farm out reading comprehension.
However grim these numbers are for AI, they’re even worse when narrowed to the sample of pre-law students:
These students use AI as often as other respondents, yet are much more skeptical of it: 66.7% say AI is acceptable for brainstorming only, against 38% of the full sample; 66.7% believe heavy reliance will leave them much less capable, against 31.6% overall. Their confidence that they could work without AI is higher (83.3% versus 62%), and their appetite for a collective AI-free semester is 100%, against 80.3% for everyone else.
“A group that uses AI often and trusts it least is not a contradiction,” the report notes, putting these figures in perspective. “Rather, perhaps it mirrors the profession’s instincts showing up early in students who appreciate traits that the law values.” That’s good news for the profession, where AI’s most responsible use case — putting aside true brute force tasks like doc review — is as a teammate overseen by a skeptical human. But it’s not great news for the frontier labs underwriting the supposed revolution.
If AI evangelists want to usher in a new era of work, they’re going to need to spend time persuading this population to trust the technology more. Right now, they’ve got some deeply unsure users who think it’s a fine tutor, but a threat to their own competence. Unless this group learns to get comfortable with a workflow that includes allowing AI to put together a first draft, it’s hard to imagine them entering the workforce capable of getting the most out of these tools even if they wanted to.
My personal experience talking to college students largely tracks this report. A lot of use… a lot of distrust. But I’d been running on vibes. Grabow lives in that world and decided it’s long past time to gather some data.
Check out the full report here.
Joe Patrice is a senior editor at Above the Law and co-host of Thinking Like A Lawyer. Feel free to email any tips, questions, or comments. Follow him on Twitter or Bluesky if you’re interested in law, politics, and a healthy dose of college sports news.
