AI is *software* - the legal professions haven't yet learned how to analyse the many limits of software - just so *alluring* to save time. Vos has drunk the coolaid so don't expect any reason hereonin from him or the AI lobbyists. A much wider discussion was warranted than has been had to get even to this point - yet here we are.
Judicial training to understand and spot use of generative AI (let along to use it themselves) will take judges out of judging so another reason for lawyers to be required to be "up front"? I am not discussing data processing and management by AI.
Who were the "distinguished Oxford academics present" who "thought that emotive decisions of that kind would be just the type of decision-making that parents would really prefer to be taken out of human hands."? I find this quite odd to say the least, and I would like to know the empirical research basis for it
Rule 4. Anyone using generative AI to produce anything be required to expressly declare on the document that they have done so and which system they used.
The difficulty is going to be the human "checking" Vos advocates - students of law and legal trainees and practising and academic lawyers would still have to maintain their own human knowledge and understanding so, really, where is the benefit of AI other than in assisting research (factual as well as legal)? And that is not "generative". Vos and Susskind (father and son) are peas in a pod. I am not a Luddite - just very careful. Law has no "reality" it is and was invented as the best means to understand and prevent (or, if necessary solve) disputes - many of which have emotional roots even if apparently concerning commercial matters
Vos is right and I think AI's case isn't helped by doomsayers who say it will replace humans. Perhaps in the distant future, maybe - but even then, only if we want it to. In the meantime it's just a tool - like word processing or e-mail. If anything there is more work than ever because of those tools!!
Our case management software has an AI tool and I've started to use it a lot. Its most magical feature is being able to ask it a question about a matter (e.g when was this claim issued? What was our last WOP offer?) and it reverts with not only largely accurate answers, but tells you which document in the file it took the answer from. Each instance saves a few minutes of frustrated scrolling and searching to find the information you want. Helpful when you have a client on the phone!!
It also has a research AI - this is still early stage and clearly will be developed more. You need some knowledge and a good nose - already I've seen it say things which weren't quite accurate. But it's a good way to quickly remind yourself of something that in the moment you can't remember, or as a jumping-off point for further research. And I'm sure it will get better.
Very true - which is why at the moment it's just a tool. The biggest worry is, if AI becomes widespread it seems to replicate a lot of what trainees do. Will it reduce the amount of training places?
AI is *software* - the legal professions haven't yet learned how to analyse the many limits of software - just so *alluring* to save time. Vos has drunk the coolaid so don't expect any reason hereonin from him or the AI lobbyists. A much wider discussion was warranted than has been had to get even to this point - yet here we are.
Perhaps one day a group of LLMs overheating in the legal ‘cloud’ will appoint a Master of the Bots…..
Judicial training to understand and spot use of generative AI (let along to use it themselves) will take judges out of judging so another reason for lawyers to be required to be "up front"? I am not discussing data processing and management by AI.
Who were the "distinguished Oxford academics present" who "thought that emotive decisions of that kind would be just the type of decision-making that parents would really prefer to be taken out of human hands."? I find this quite odd to say the least, and I would like to know the empirical research basis for it
If anyone who was present (Pembroke, 27 November) can let me know I'll be happy to report what they say.
Rule 4. Anyone using generative AI to produce anything be required to expressly declare on the document that they have done so and which system they used.
The difficulty is going to be the human "checking" Vos advocates - students of law and legal trainees and practising and academic lawyers would still have to maintain their own human knowledge and understanding so, really, where is the benefit of AI other than in assisting research (factual as well as legal)? And that is not "generative". Vos and Susskind (father and son) are peas in a pod. I am not a Luddite - just very careful. Law has no "reality" it is and was invented as the best means to understand and prevent (or, if necessary solve) disputes - many of which have emotional roots even if apparently concerning commercial matters
Thanks for this thoughtful comment.
Another concern is training — not just for lawyers, who can be expected to get on with it, but also for judges, whose training needs to be provided by the Judicial College: https://www.judiciary.uk/about-the-judiciary/training-support/judicial-college/
Vos is right and I think AI's case isn't helped by doomsayers who say it will replace humans. Perhaps in the distant future, maybe - but even then, only if we want it to. In the meantime it's just a tool - like word processing or e-mail. If anything there is more work than ever because of those tools!!
Our case management software has an AI tool and I've started to use it a lot. Its most magical feature is being able to ask it a question about a matter (e.g when was this claim issued? What was our last WOP offer?) and it reverts with not only largely accurate answers, but tells you which document in the file it took the answer from. Each instance saves a few minutes of frustrated scrolling and searching to find the information you want. Helpful when you have a client on the phone!!
It also has a research AI - this is still early stage and clearly will be developed more. You need some knowledge and a good nose - already I've seen it say things which weren't quite accurate. But it's a good way to quickly remind yourself of something that in the moment you can't remember, or as a jumping-off point for further research. And I'm sure it will get better.
But just like
... but just like humans it can make mistakes?
Very true - which is why at the moment it's just a tool. The biggest worry is, if AI becomes widespread it seems to replicate a lot of what trainees do. Will it reduce the amount of training places?