It is recommended to create an operational framework accompanied by general ethical guidelines that emphasize principles such as data privacy and confidentiality.
Developing a document to standardize the use of AI within the law firm can be useful, including the creation of a section dedicated to 'prompt engineering'.
Although generative artificial intelligence such as ChatGPT is still in its infancy, many law firms are considering the use of AI. According to a report by Juro, a provider of technology solutions for the legal world, 64% of lawyers say their law firm is already using or has considered using AI.
However, certain guidelines should be considered before implementing AI in a law firm. Neosmart offers several recommendations.
How to start using AI in a law firm
In general, it is advisable to approach AI gradually to understand its value and integrate it into the law firm. This can start by using it for less critical tasks, such as assisting with document review or even creating email text for internal communications. Then we can move on to extracting key data from a contract or summarizing texts for the client in plain language. Later, we would use functions that involve rewriting legal texts or generating contract clauses.
The timeline depends on the law firm's interest in introducing AI, its resources, and its needs. However, certain aspects must be clarified before the final introduction of AI. One of the most important is data protection. Some AI models, such as the one used by ChatGPT, are trained with information entered by users. This means that they do not record this knowledge like a conventional database, but at a level that could lead to complications.
It is obvious that some of the information processed by a lawyer is confidential and must not be leaked to the outside world. This could happen indirectly, as the AI application could pass on training data to another user. While some services, such as ChatGPT itself, have options to protect personal data, it will always make sense to anonymize information so that no proper names or other identifiers are present.
In addition, it is advisable to read up on the terms of use and privacy policies of the AI applications we use. Those that are exclusively focused on the legal field tend to have confidentiality requirements in mind. Counsel by Casetext, for example, claims not to use users' data and to keep it private.
Establishing a basic ethical guideline
With the use of AI, new debates will arise within law firms about how it should be used and for what purposes. One recommendation is to establish basic ethical guidelines to consider. This could be a simple document with a few summarized points that give an idea of the necessary ethical framework for the use of the tools.
Some of these ethical principles are obvious, such as the need to protect privacy and data protection, which we have already discussed. Another point is transparency about what has and has not been done with AI within the company. It should also be clarified who is responsible for the work. After all, AI is a machine, and the people who check the results should take responsibility for them.
How to use AI efficiently
Once the framework in which AI can be used in the law firm is clear, the next step is to optimize the use of the technology. Three initiatives will improve lawyers' productivity when using the new tools.
Introduction of a standardized methodology for the use of AI
The introduction of common procedures for the use of AI will be of benefit to the firm's lawyers. The aim is to define the best ways to use the technology, depending on the firm's needs and resources. A dynamic guide could be created that evolves as the applications are used. The idea is that everyone can contribute to this document with their experience.
One of the most important sections of this guide will deal with optimizing 'prompt engineering', an area of improving communication with AI applications to achieve better results. Guidelines will be provided to get the best out of the AI models. Here, too, the entire team could contribute with its experience.
A training program
A training program for the use of artificial intelligence has its advantages. Professionals do not have to learn everything on the first go or rely on a document to work with AI. They can acquire basic practical knowledge through training days or an online course.
A system for reporting incidents
Just as important as optimizing communication with AI is reporting incidents that occur. If you know the weaknesses, inaccuracies, and errors of artificial intelligence applications, you can use them more consciously. It therefore makes sense to set up a channel for reporting such errors to alert the rest of the team and try to avoid or improve them where possible.
