I would like to create a prototype with davinci or one of the other models that would scan legal documents for poor people who cannot afford a lawyer. Then creating an app that would help those people through blockchain technologies and openai models. I think it would be a cool way to make the world a better place. If we could create it.
That’s a cool idea!
How do you want to help people? The first thing I thought of was a program used for translating difficult legal concepts from documents into a summery while defining specific verbiage for those who are unfamiliar in the field. Is that close to what you were thinking?
That is great input. I think that it is very close to what I was thinking as well. Also looking up past cases that might be pertinent to their case. Then let them know what defense or arguments would be the most likely to win based on past cases. So I guess we would start with a document crawler that summarizes legal concepts to a 4th grade level of reading?
What exactly do you have in mind with blockchain? Because to me NLLP + Blockchain could be about AI writing smart contracts.
I am currently playing with constitutional law (for my country or origin). However, I am thinking about Constitution + Decisions, and something to translate legal jargon (this translation has been done already, somewhere but can’t remember his twitter account).
The reason I would want to add blockchain is so that it’s easier to look up past cases we’ve helped. Plus you could fund it with coin exchanges and smart contracts. So even if a government doesn’t want someone to have access to the app. It would be decentralized and accessible from anywhere.
My humble advise to you is to narrow the scope of whatever your goal is. Law is too broad. Think about three most important justice systems where you can make a change, then choose one. These are:
- Civil and Commercial Justice
- Labour Justice
- Criminal Justice
It has been observed that some countries have needs depending on their social political and economical circumstances.
Degree → Needs
Developed countries → Civil and Commercial
Somewhat developed countries-> Labour
Third world countries-> Penal and criminal proceedings
Estonia and China have very advanced AI systems in place already: AI judges.
Countries and AI language models
There are a number of work in progress, and they have a human in the process (meaning after a certain point one human take control of the interaction).
it may take time until day to day legal challenges reach the constitution level
Time is what we have.
I am not thinking about (constitutional) debating AI (some work by IBM already here). I am thinking about tools to help understand and give an approximate interpretation of our fundamental rights.
I am adept to the living tree doctrine of constitutional interpretation that says that a constitution is organic and must be read in a broad and progressive manner so as to adapt it to the changing times.
One problem with the legal practice is we trust too much in our lawyers and the whole system, everywhere, and it is a system subject to failures and corrections. We can help them (lawyers and those in the system), and help ourselves by reading. We are part of the problem. We can have our interpretation of the law with the help of AI, and the lawyer, and then put the lawyering machine to work in the defense of our rights.
I’m interested in this project and can help! Got a contact form or anything and we can get in touch to discuss details?
Trying to integrate a blockchain based backend will probably kill the project at this stage.
Plus if the end user = person with little financial resources for legal help…
…then the end user is not likely concerned if the app uses a decentralized ledger or not anyways.
[my previous exp includes: once built a popular DApp on Ethereum]
You can send me your contact info in a DM. I will get back to you within a week or so. I am thinking of holding remote meetings on Fridays to see where we are with the project and what we would need to take as next steps.
So I found a couple of different types of lawyer ai that exist now. I think if we were to model some of it after different ones we would get the best results. The IBM debater would be a good outline for ai lawyer if you needed it to represent you in court.
Aside from that, I found DoNotPay.com which focuses on parking tickets and other things. But that one you have to pay in order to even get access to it.
#psychologyoflaw + #naturallanguagecognitivearchitecture =ΛI PSYCHOLOGY™ ?
Great APP! I got a brandnew and updated fridge by using it
Hi, I am a capital markets lawyer by training with expertise in corporate law, M&A, securities, corporate governance and ESG. There are many legal technology firms working on making laws more accessible to everyday people, broadly termed A2J or access to justice. Two key A2J areas that are being disrupted, and that could benefit from creative, GPT-3-driven solutions, are landlord-tenant law and divorce law, If you don’t have the legal background yourself, it goes without saying that you need to bring experts in those fields on to your team. If you try to focus on “law” generally, I fear you will drown, unless you partner with a large company like Thomson Reuters.
Also, I forgot to mention, for what it’s worth, I don’t see any value add from blockchain in A2J use cases.
I agree with everything @lmccallum says above advising that OP get lawyers on the team and focus, but I do want to point out one reason to be optimistic about the opportunity for blockchain to add value in landlord-tenant and divorce law. From what little I know about L/T and divorce, both are frequently characterized by highly disputed facts and allocation of responsibilities. Blockchain’s ability to create immutable, transparent records ought to be helpful in building the record for prospective litigants.
Who is ever interested in starting this project should know it will be open-sourced and probably version zero will include helping people with apartment laws, and litigation requirements to sue someone for small claims. Since most lawyers can’t help with small claims cases anyway depending on the state.
I run Docket Alarm, a litigation data company. We’d be happy to supply sample data to this project, with live litigation data direct from the courts. Here are some sample searches:
- 9M Evictions: Search court records on tag:(eviction or landlord)
- 20M Debt / Collections: Search court records on tag:(debt or collection)
Besides the history, there are live case feeds to many courts. There are some interesting A2J problems that can be solved with proper alerting / deadline management (e.g., automatically draft a document the moment it is required).
Happy to chat more about this.