What’s the best gen AI model for legal professionals? According to one article, the answer is – not surprisingly – a very “legal-ish” response.

Of course, to know the best gen AI model for legal professionals, you first need to know what the gen AI models are that legal professionals should know. And this infographic, courtesy of Cassandre Coyer and Rhys Dipshan of Legaltech^® News, identifies six of them:

1. LLaMA (Large Language Model Meta AI), developed by Meta
2. GPT (Generative Pre-trained Transformers, which include GPT-3.5, GPT-4 & GPT-4 Turbo), developed by OpenAI
3. LaMDA (LaMDA: Language Models for Dialog Applications), developed by Google (albeit, no new versions since January 2022)
4. PaLM 2 (Pathways Language Model), developed by Google
5. Claude 2, developed by Anthropic
6. Gemini, developed by Google (announced last month)

Cool infographic. And it accompanies this article (What’s the Best Gen AI Model for Legal Professionals?) by Rhys Dipshan (also on LTN).

So, what is the best gen AI model for legal professionals? You probably shouldn’t be surprised that the answer is very legal-ish: “it depends”.

The article discusses several of the different alternatives and quotes experts, who have identified which models are best suited for specific tasks. For most of the tasks, the answer is either GPT-4 or Claude 2, though the article notes that “Meta’s Llama 2—is open source, and is meant to be changed and tailored to the whims of a software developer”. But the article also states this:

“There’s no easy answer, therefore, to what model is the best for legal tasks. It’s dependent on a variety of specific factors and contexts, leaving legal professionals with the time-consuming, though likely rewarding, work of testing out and fine-tuning different models for different tasks.”

Surprised? No? Me either.

Given that there’s one highly publicized study that just came out from Stanford University that states that large language models produce hallucinations between 69% and 88% of the time when queried about a legal matter (but included GPT 3.5 instead of GPT-4 in the study, just sayin’), it seems wise to not put all your eggs into a single basket when it comes to the use of gen AI models for legal.

The infographic and the article are very informative and you should definitely check them out. But the best approach for many use cases (not just legal) is one that can be flexible to use the right gen AI model for the use case. Because the answer to the question “what is the best gen AI model for legal professionals?” is “it depends” and it will be for the foreseeable future.

So, what do you think? What gen AI model(s) does your organization use? Please share any comments you might have or if you’d like to know more about a particular topic.

Image created using GPT-4’s Image Creator Powered by DALL-E, using the term “robot juggling balls”.

Disclaimer: The views represented herein are exclusively the views of the authors and speakers themselves, and do not necessarily represent the views held by my employer, my partners or my clients. eDiscovery Today is made available solely for educational purposes to provide general information about general eDiscovery principles and not to provide specific legal advice applicable to any particular circumstance. eDiscovery Today should not be used as a substitute for competent legal advice from a lawyer you have retained and who has agreed to represent you.