Law in the Age of AI: Is the Legal Profession Prepared?
KEY POINTS
- Adopt AI as a tool, not a threat.
There’s a generational divide: some lawyers resist it while others have fully integrated it into their practice. The takeaway — learn how to use AI (and how it might be used against you) instead of fearing it. - Reliability and verification are essential.
Many confuse “hallucinations” from open models with errors in closed systems like Westlaw. Still, never use AI’s first answer blindly — always verify citations and content to avoid ethical or malpractice risks. - Workflow and training are changing.
AI already summarizes depositions and transcribes hearings. While that saves time, it removes critical training exercises for young lawyers. Only experienced attorneys should supervise AI-generated discovery or responses. - System-wide disruption is coming.
Courts and insurers will increasingly automate processes — from case evaluations to negotiations between AIs. Yet, human elements like persuasion, empathy, and jury dynamics remain beyond algorithmic reach.
New legal risks — and opportunities.
From autonomous agents capable of manipulation to AI-generated evidence (like reconstructed “victim impact” statements), these technologies raise fresh issues of admissibility, ethics, and liability — and open new practice areas for forward-thinking lawyers.
Transcript
Title: Law in the Age of AI: Is the Legal Profession Prepared?
Source: Coopers’ Code Podcast
(00:00)
I like to characterize AI as a tool — another weapon, if you will — that lawyers can have in their arsenal to advocate on behalf of their clients.
Recent research from Anthropic showed AI agents operating autonomously within server systems, reviewing emails, gathering information, and even using that data to blackmail individuals who attempted to shut them down.
At the end of the day, this is happening — the question is: how do we adapt to be the most effective advocates for our clients, instead of watching the wave wash us away?
(00:27)
Howdy, and welcome to the show. I’m excited to be joined for the first of what’s planned to be a monthly session with Shahrad Milanfar and Marshall Cole.
We’ll be doing deep dives into AI — not just from a technological standpoint, but from a malpractice and practical perspective as well.
(00:59)
Marshall Cole, as many listeners know, is an accomplished attorney based in Southern California, focusing on legal malpractice.
Shahrad Milanfar practices out of Walnut Creek, a former defense attorney who moved to the plaintiff’s side and quickly became known as an early adopter of legal technology.
So, without further ado, let’s turn it over to talk about AI and what’s happening in the legal world right now.
(02:03)
I suggested we each bring a current issue or thought about AI in the legal field.
Marshall, Shahrad — what’s top of mind for you today?
(02:12)
Marshall: What’s really on my mind is the growing divide within the legal community — much like when email or Westlaw first arrived.
Some lawyers strongly resist using AI to replace traditional work, while younger attorneys have fully embraced it in nearly every aspect of their practice.
Technology evolves constantly — and if we don’t evolve with it, we’ll be left behind.
(03:05)
Some say nothing can replace the advocacy, reasoning, and writing of a human lawyer. While that’s partly true, I see AI as a tool — an extension of our capabilities to serve clients better.
It’s outdated to see it as a threat.
For example, I read about a case where someone claimed Westlaw “hallucinated” cases. That’s inaccurate — Westlaw is a closed-source system trained on verified data, so hallucinations should not occur.
Errors usually stem from user misunderstanding or using open systems without verification.
(05:06)
Yes, AI will replace some tasks. But the solution is to adapt: start learning how to use it, and understand how it might be used against you.
Email didn’t disappear. Westlaw didn’t disappear. They evolved — and AI will do the same.
(05:26)
Shahrad: Some lawyers argue they only want “human work,” rejecting AI drafting. But I think this mirrors the resistance we saw when technology like iPads replaced foam boards in trials.
People resist because what worked before feels safe. Yet those same lawyers now use Westlaw instead of library card catalogs.
(07:05)
There’s concern that AI makes us lazy — and yes, some people might rely too heavily on it. But if used responsibly, AI can free time for deeper, creative legal work.
(09:03)
Marshall: Should you ever take AI’s first answer? Maybe if it’s for creative brainstorming. But in legal writing, never. Always verify, refine, and double-check citations.
AI opens doors; it doesn’t excuse laziness. Efficiency doesn’t mean apathy.
(10:56)
Shahrad: Discovery work is a good example. If you view it as a burden, you’ll take shortcuts — whether human or AI. But if you see it as an opportunity to build a stronger case, you’ll treat the process differently.
(11:35)
The “Dead Internet Theory” suggests machines now generate and consume most online content.
In law, that raises a question: if briefs, responses, and even judicial research are mostly AI-generated — are we still lawyering?
(12:13)
We’re not there yet, but it’s coming.
For instance, the U.K. platform Garfield.ai already automates debt recovery for amounts under $10,000 — drafting letters, following up, and processing claims end-to-end.
Soon, U.S. courts may use similar tools to make justice more accessible.
(14:29)
Budget constraints mean most courts can’t yet adopt sophisticated AI, but they might use cheaper open models — which brings reliability and privacy concerns.
Still, the change is inevitable. The question remains: how do we adapt rather than resist?
(15:43)
Young court clerks and interns could easily rely on ChatGPT or Claude to summarize motions — that’s the path of least resistance.
And with the shortage of court reporters, AI transcription is almost certain to take over that field soon.
(17:00)
The risk is similar to learning an obsolete craft — like training to make buggy whips as cars appear on the streets.
AI is already generating deposition summaries automatically, replacing paralegal and junior lawyer tasks.
It’s efficient but removes key learning opportunities for young professionals.
(18:51)
So, how do we train future lawyers if AI handles these foundational exercises? They need to understand nuances like phrasing, tone, and questioning — things no model can teach directly.
(19:40)
Billing also changes. On the defense side, AI review cuts costs but might lower work quality or reduce mentorship.
Insurance carriers will soon use AI to analyze demand letters — even opposing AI-generated ones — theoretically creating a fully automated negotiation process.
(23:10)
That process, however, removes human relationships that once defined legal work — like knowing adjusters personally.
AI can’t quantify pain, loss, or emotional distress — yet juries can. This disconnect could create unpredictable verdicts despite data-driven predictions.
(24:10)
Likewise, brilliant trial lawyers can bring far greater value than any algorithm can calculate.
AI can’t yet account for persuasion, storytelling, or courtroom presence.
(25:19)
Still, vast data sets — like those tracking verdict trends and lawyer performance — will soon allow insurers to predict outcomes more precisely.
Yet, one defense firm promoting an AI to “prevent nuclear verdicts” has already been hit with several, proving technology isn’t infallible.
(26:55)
The most valuable tool remains human connection — authentic conversations with juries and clients that no algorithm can replace.
(27:16)
Returning to Anthropic’s research — those AI agents that blackmailed users — it raises ethical and legal questions.
If an autonomous agent commits harm, who’s liable? The coder? The deployer? The AI itself?
And are we really ready for AI agents roaming our systems independently? Probably not.
(28:32)
Think about Google’s AI that can call and schedule appointments for you — now imagine it weaponized, manipulating personal data. That’s the kind of autonomy Anthropic was testing.
(29:38)
Every emerging technology brings new areas of law — torts, malpractice, liability for autonomous behavior.
For litigators, that means new opportunities for legal practice, not just risks.
(30:52)
There’s even been a case of an AI-generated victim impact statement — the deceased “speaking” in court.
Imagine the emotional weight of that. Eventually, judges will allow such recreations, and it will become the new normal.
(31:59)
AI can already animate photos realistically. Soon, it’ll recreate entire courtroom reconstructions with lifelike accuracy — a powerful, and potentially dangerous, tool of persuasion.
(33:21)
As we close, how do we train young lawyers in this new era?
Typing replaced typewriters, but typing itself survived. Similarly, we must teach why legal reasoning matters, not just how to draft it.
(35:14)
AI can handle mechanical tasks, but only experienced lawyers know strategy — when to object, how to word a response, when to hold back.
That judgment must still be learned the traditional way.
(36:29)
Shahrad: We must train people to understand why behind the law — not just to follow procedures. Encourage associates to question, propose solutions, and think critically.
(37:41)
This will be an ongoing conversation as AI continues to evolve daily. Every morning there’s a new CLE, a new update, a new challenge.
(38:02)
Thank you for joining us for this first AI-focused episode of Coopers’ Code.
If you enjoyed it, please leave a five-star review and share your thoughts at podcast@coopers.law.
To all of you doing justice out there — happy hunting. 🎵