AI in law: Balancing performance, risks, and professional responsibility

In a discussion that had a prominent focus on the experimental use of such technologies, it was evident that firms are monitoring AI performance and identifying the risks associated with the technology ultimately taking over tasks in the legal profession that were once conducted by humans.

In a recent survey conducted by Lawyers Weekly, it was discovered that AI is already reshaping BigLaw graduate training by automating routine tasks and giving young lawyers initial exposure to complex legal work, client interaction, strategic thinking, and higher-value responsibilities.

With aspiring lawyers having access to these tools throughout their studies, concerns are often raised surrounding the risk of AI and whether it may replace the key traditional obligations of a young lawyer.

Moreover, lawyers – such as Golab – have identified major risks associated with client data, security, governance, and professional responsibility.

“You really must ensure you’re using multifactor and that you fully understand that you’re using, you’re not using your home licence for an AI tool, that you’re using your work licence and that you’ve properly logged into that system so that there is auditability and visibility,” Golab said.

“You’re effectively duplicating data where you’re loading data into the tool for a task, and so you have the headache of, how can you safely delete, destroy that data or bring it all back into the document management system if that is your permanent storage?”

VIEW ALL

Further to these risks, Golab indicated that AI agents are becoming increasingly integrated into legal practice, warning that legal practices should diligently evaluate information security, indemnity, and ethical obligations prior to the implementation of such systems within practical legal environments.

“You really must not pilot something with real-life client data … and it’s very difficult to pilot something with synthetic data that isn’t realistic,” Golab said.

“But you have to be very careful in ensuring [that] if you’re doing a pilot, that you have some form of agreement or contract in place simply for the protection of your data.

“And then there’s the liability and the risk side of, are you ultimately responsible for the actions of this agent? That would definitely be the primary one.

“At the end of the day, you are responsible for your work.”

Additionally, the discussion showed that firms need to evaluate AI based on measurable results rather than essentially assuming that new technology will enhance their work.

These discussions have arisen as technology companies Thomson Reuters and Laurel have partnered to assist law firms with measuring AI’s impact by linking usage to efficiency, profitability, matter performance and ROI.

“You really need to have the discipline of having pilots and measuring and fully assessing what the tool does versus your perception of and what you’re told that tool does,” Golab said.

“There are enormous measurement tools that we have simply because these are very large-scale systems that are tracking every decision a lawyer makes, and so we have a very high precision in measuring tools.

“Everyone relates to a painful multi-day task, and then to be able to measure it and to be able to demonstrate that you took that task that may have been a 20-hour task and you’re doing it in five hours or something.

“So, definitely, measurement, very much education and communicating this in practice groups and reinforcing matters.

“Having ways to measure things is super important to have a very good idea about how every day is different, you’re always doing different tasks, but you generally have a pretty good ballpark of within certain parameters.”