This is an area for further development and thought...
If a LLM can pass the bar, and has a corpus of legal work instantly accessible, what prevents the deployment of the LLM (or other AI structure) to provide legitimate legal services?
If the AI is providing legal services, how do we assign responsibility for the work (to the AI, or to its owner)? How to insure the work for Errors and Omissions?
More practically, if willing to take on responsibility for yourself, is the use of AI going to save you money?
A human that screws up either too often or too spectacularly can be disbarred, even if they passed the bar. They can also be sued. If a GPT screws up, it could in theory be disbarred. But you can't sue it for damages, and you can't tell whether the same model under a different name is the next legal GPT you consult.
> If a LLM can pass the bar, and has a corpus of legal work instantly accessible, what prevents the deployment of the LLM (or other AI structure) to provide legitimate legal services?
The law, which you can bet will be used with full force to prevent such systems from upsetting the (obscenely profitable) status quo.
Re your first point: it's not conscious. It has no understanding. It's perfectly possible the model could successfully answer an exam question but fail to reach the same or similar conclusion when it has to reason it's own way there based on information provided.
Great point, LLM will not be great at ground breaking law.... But most lawyers aren't. That's to say, most law isn't cutting edge. The law is mostly a day-to-day administrative matter
Careful, there are plenty of True Believers on this website who really think that these "guess the next word" machines really do have consciousness and understanding.
The obvious intermediate step is that you add an actual expert into the workflow, in terms of using LLMs for this purpose.
Basically, add a "validate" step. So, you'd first chat with the LLM, create conclusions, then vet those conclusions with an expert specially trained to be skeptical of LLM generated content.
I would be shocked if there aren't law agencies that aren't already doing something exactly like this.
What if they were liable? Say the company that offers the LLM lawyer is liable. Would that make this feasible? In terms of being convincingly wrong, it's not like lawyers never make mistakes...
You'd require them to carry liability insurance (this is usually true for meat lawyers as well), which basically punts the problem up to "how good do they have to be to convince an insurer to offer them an appropriate amount of insurance at a price that leaves the service economically viable?"
Given orders of magnitude better cost efficiency, they will have plenty of funds to lure in any insurance firm in existence. And then replace insurance firms too.
"In terms of being convincingly wrong, it's not like lawyers never make mistakes..."
They have malpractice insurance, they can potentially defend their position if later sued, and most importantly they have the benefit of appeal to authority image/perception.
All right, what if legal GPTs had to carry malpractice insurance? Either they give good advice, or the insurance rates will drive them out of business.
I guess you'd have to have some way of knowing that the "malpractice insurance ID" that the GPT gave you at the start of the session was in fact valid, and with an insurance company that had the resources to actually cover if needed...
Weirdly HN is full of anti AI people who just refuses to discuss the point that is being discussed and goes into all the same argument of wrong answer that they got some time. And then they present anecdotal evidence as truth, while there is no clear evidence if AI lawyer has more or less chance to be wrong than human. Surely AI could remember more and has been shown to clear bar examination.
"while there is no clear evidence if AI lawyer has more or less chance to be wrong than human."
In the tests they are shown to be pretty close. The point I made wasn't about more mistakes, but about other factors influencing liability and how it would be worse for AI than humans at this point.
If a LLM can pass the bar, and has a corpus of legal work instantly accessible, what prevents the deployment of the LLM (or other AI structure) to provide legitimate legal services?
If the AI is providing legal services, how do we assign responsibility for the work (to the AI, or to its owner)? How to insure the work for Errors and Omissions?
More practically, if willing to take on responsibility for yourself, is the use of AI going to save you money?