As you perhaps know, the developer of DoNotPay recently attempted to bring a robot lawyer into a physical courtroom. He was rebuffed. But will we soon have robot lawyers in the courtroom in the next year or two in your view?
Despite the fact that AI has hit an inflection point recently, it’s unlikely that we will see AI assistance directly interacting in the next year. However, in the next two or three years I think it is highly possible select jurisdictions will test it. The reason for the delayed acceptance is twofold. Attorneys and the overall legal community, including the judicial system, tend to be extraordinarily risk adverse. They have also created a guild which is very protective of its craft and people. Combining the two and the idea that AI will act as a lawyer in the court room imminently is doubtful.
That all said, a good percentage of prep work that litigators perform leverages AI behind the scenes in their research, and increasingly strategy. Legal outcomes can now be empirically weighed via prediction models using similar natured, previously litigated cases, and their docket information by judge and jurisdiction. This means that judges have patterns which can be extrapolated via AI, enabling strategy by savvy attorneys. The logical next step is the presentation of facts and arguments, which is increasingly becoming feasible.
Is it only a matter of time before large law firms cease to exist, in your view? If they do persist, will the nature of their work have changed?
I believe large swaths of legal jobs will be eliminated in the coming years for all firms. Large law firms should be able to weather the storm better than most medium and small firms that don’t have a very specific niche. Large law firms will survive by handling highly complex issues. Across the industry, it’s the cookie cutter work that most firms do now, which will implode. I’ve been advising law firms on the pending technological changes afoot which will vastly reduce the need for routine lawyerly work. AI search, workflow, and automation tools combined with the recent large language model (LLM), designed generative AI, and natural language processing (NLP) models will birth more consumer and business to business websites that automate many aspects of the transactional practice. Conversely, in litigation it is conceivable that a Kleros (decentralized alternative dispute resolution) system could be a model to resolve conflict rather than leveraging the courts.
You have also written that blockchain technology has potential to disrupt the legal sector. Will we soon see “smart contracts” replacing many lawyer written commercial contracts?
Absolutely, in the near future, many commercial contracts will be written as “smart contracts”. I will take that one step further, NFTs (Non Fungible Tokens) will be the backbone of asset ownership, e.g. deed to your car or home. It will be way more trackable and transferable.
Per the question about smart contracts, I authored this example of a contract which I think illustrates the future state of blockchain based legal agreements.
Legal Smart Contract:
You are an attorney writing a Trust for your client. The Trust is simple and a tad ridiculous in this example. It stipulates that upon the parent’s death, their two kids must be married in order for them to split the estate. If one kid is married and the other kid is not, the kid that is married gets the entire estate. For simplicity the assets are all liquid in this example.
Where blockchain can improve this…
The Trust is written and then codified. The code identifies the parameters that are contingencies or possibly subject to change. Saved as a smart contract on a blockchain, it is now in an immutable state but has actionable items imbedded in it. The only people that have access to this document are the attorney that drew it up and her client. Once it is on the blockchain, the smart contract – with the coded parameters - starts checking every day through a trusted source, called an oracle (affirmed public record), to see if both parents are alive. One day the computer identifies that the parents have passed. Heartless, the computer jumps to the next task to determine if both kids are married. Through another API computer call to that oracle, it finds out that one kid is married, and the other kid is not, and subsequently sends 100% of the liquid assets to the kid that is married – into their digital wallet. This is a self-executing smart contract on a blockchain where in the future state, no human (lawyer) intervention is needed.
Other ways that blockchain (possibly in tandem with AI) might further “democratize” legal services?
There are a handful of services akin to ChatGPT either out or coming out in the coming months. The humanlike interaction of the chat will enable people without the resources to get better advice on legal issues than ever before. I anticipate full catalogs of legal decisions combined with large language model (LLM) designed to use generative AI and natural language processing (NLP) to vastly improve access to justice and foster new LegalTech businesses that will offer up legal workflow at fractions of the cost of hiring an attorney today. A very high percentage of legal work can be considered repetitive on the transactional side.
DAOs – could impact the way that businesses are formed, and lawyers will have to understand this shift. This too could automate the way that companies are created and run.
Are there concerns or dangers, though, in a world without lawyers?
Yes, no question there is significant risks in the early days under the use of both AI and blockchain, i.e. trusting in the code. In the early phases of any technology, especially in the legal industry, mistakes are not acceptable. To that end, this community will wade into the waters with less significant cases, e.g. parking tickets, before the more impactful issues, but those weighty cases will eventually come, once the others have been proven to work well. Bias will continue to come up as a real concern, but as long as it does, it will keep people focused on keeping it in check.
Will we ever be able to trust algorithms with the most complex legal cases, like sentencing decisions in capital cases?
I firmly believe in 15-20 years, we will trust algorithms to adjudicate the most complex legal cases. My rationale is thus. I believe more and more contracts will rely on code and increasingly become more universal. There are communities working on a universal coding language for contracts which can be better understood by machines. Over time that code will be more trustworthy, defined, and clear. Ultimately, many matters could be weighed compared to other similar cases and possibly judged by a machine. Additionally, with an increase reliance on digital data of every type, the availability of provable data will also increase. At the very least, these algorithms will be a sort of augmented intelligence for judges to help them make a decision.
Unquestionably, the legal industry is primed to be significantly impacted by both AI and blockchain in the not-too-distant future.