Author: Gloria Qiao, Founder & CEO of Trusli
We have written a lot about AI and law. We have discussed the challenges the clients have while trying to find lawyers and how we solved them. We also discussed how to manage legal practices using software and AI in depth. When discussing artificial intelligence and law, we often get asked: are you saying AI can replace lawyers? Do you mean that documents will draft themselves? When we were working in self-driving cars at Zoox and Lyft, we also got asked: can cars drive themselves? The answer is yes and no. Yes, maybe eventually. No, not in the near future. So how does AI in law work? What are the best use cases of AI for lawyers? Here are some thoughts.
Where AI Excels
In our opinion, there are certain tasks in which AI can significantly outperform humans.
AI has a tremendous capacity to gather a large amount of data and analyze them based on certain predefined criteria. AI can tap into databases accurately and efficiently with very little room for error at a tremendous speed. This technology does not forget, does not make human errors in mislabeling such data, and does not get tired. When the data is structured, the ability of AI to gather, retain and analyze such data far exceeds human capabilities.
Pre-Defined, Repeatable Tasks
AI can perform predefined tasks that are simple and repeatable without deviation and errors. If we can define the tasks clearly for the AI (e.g., follow the lane and drive, and stop when you see an obstacle), then AI will perform per instruction without any deviation or mistakes. Similar to the last point, AI will make no mistake within the pre-defined parameters. AI does not get bored or mind the tediousness of the tasks.
Related to the last point, AI is excellent at creating standardized processes and automation. Once you teach it the entire workflow and set up the end-to-end process, AI will follow through and ensure the process is performed with consistency and automation. Finally, with enough learning, the AI can also predict what the user will need or do next and suggest these actions intelligently.
Where AI Struggles
However, even though there are talks of AI being self-aware and even general artificial intelligence, at this point, artificial intelligence is not capable of perceiving, understanding, and reacting to things that it has not been taught. Simply put, if you expect an “artificial intelligence lawyer” that can replace actual lawyers in actual negotiating and drafting full-blown legal documents, we are still very, very far away from that stage.
At this moment, AI cannot perform certain tasks that are complex and need to weigh the pros and cons of various options and make a conscientious choice. It cannot think outside the box, create a strategy, decide on trade-offs, and weigh complex factors to make a “call.” It’s also not great at being creative and inventing something that does not exist yet. Finally, AI, of course, cannot replace humans in terms of empathy and emotion, and it cannot communicate like an actual human being.
Where Can We Best Use AI for Attorneys?
We’ve discussed the general trends in legal tech before, but what are the best use cases for AI in law and AI for attorneys?
Search (Legal Research, E-discovery, Due Diligence, Platforms for Finding a Lawyer)
AI can retain and categorize a large amount of legal data, making legal search simple and easy. Ross Intelligence built an excellent research platform where you can search all case laws by typing in a few simple search phrases or sentences. Too bad it closed due to a lawsuit with Thomson Reuters. Another use case in this search area is E-discovery, pioneered by companies such as Onna and Disco. We have yet to see any leading platforms capable of matching clients with different types of lawyers using AI. Trusli has a patent on this application and will be the pioneer in this area.
Process Automation (CLM, Case Management, Complaint Filing)
Any young lawyer who has worked at big law firms knows how tedious legal work can be. There are many processes; saving, naming, searching for contracts, and exchanging redlines with opposing counsel to name a few. Many of these processes can now be automated using AI.
Ironclad, along with other notable vendors (e.g., Linksquares), is an excellent Contract Lifecycle Management platform. Similarly, several case management software solutions use AI for litigation, such as Litify. Additionally, AI is excellent at filling certain forms following predefined, simple steps. DonotPay is the pioneer in this area. For simple notarization, Notarize is also very good and efficient.
At this point, no one has built a comprehensive, end-to-end legal practice management platform powered by AI. Trusli is also a pioneer in this area. Our vision is that we will gather user-specific data and behavior, learn from them, and then automate a custom-built workflow for the user. More importantly, the Trusli legal AI will also suggest the following steps based on prior learning or behavior, further taking the AI for lawyers to the next intelligence level.
Simple Form Generation and Drafting (Form Contracts, Wills & Trusts)
To our point about simple and repeatable tasks, if you only need to draft a simple contract that can be repeated from previous forms, AI is an excellent way to go about it. For example, ContractpodAi and LawGeex attempt to use AI to help lawyers draft contracts. However, if the contract becomes complex, AI may or may not be able to draft it effectively. See our point below. Similarly, there are certain types of legal documents that tend to follow a predefined form, such as wills and trusts. Farewill is a good example of this use case.
What Are the Worse Use Cases of AI for Lawyers?
When it comes to artificial intelligence and law, to our point above, “AI lawyers” is less effective in areas where the issues are complex, and one needs to weigh multiple factors and make trade-offs. Currently, AI is not great at drafting complex contracts with numerous factors, and one needs to make trade-offs and negotiate with the counterparty. Similarly, because of the complexity and multi-factor nature of legal contracts, it will be extremely difficult for AI to come up with strategies to maximize the clients’ benefits and achieve their overall goals, as it will be almost impossible to quantify and set parameters for all this and teach the AI how to consider and evaluate all of this fully.
Ultimately AI is a machine. At this point, legal AI cannot replace a human lawyer’s capacity to listen emphatically to clients and provide them with human communication that helps them understand their issues and assure them that they are in good hands.
The bottom line: AI in law is complex. The legal industry needs to adopt AI to replace human lawyers in areas where the tasks are data-heavy, repeatable, and automatable. However, we still need human lawyers for complex contractual and litigation issues, strategy and creativity, and empathic, communicative counseling that all of us want and need. The “artificial intelligence lawyer” is not here yet, just as fully autonomous vehicles with no human supervision or guidance. Artificial intelligence and law is an ongoing journey.
Check out how we’re using AI to automate a firm's end-to-end practice management at Trusli.