Extract from Rhys Dipshan’s article “What’s the Best Gen AI Model for Legal Professionals?”
While generative AI exploded into the mainstream in late 2022, generative AI models are certainly not new—AI developer OpenAI, for instance, released a number of versions of its GPT model series years before announcing GPT-3.5, which powers the free version of ChatGPT, in November 2022.
What has changed over the past year, however, is the speed at which new models are entering the market and updated versions are coming online. For legal professionals, it can be difficult to know which models can best bring efficiencies to their day-to-day jobs, specific legal tasks and the overall delivery of legal services.
Fortunately, some are hard at work testing these generative AI models on a range of legal tasks. While there are obvious differences in each model’s performance with legal work overall, it’s becoming evident that comparing them isn’t a clear-cut endeavor. After all, the requirements for specific legal tasks very widely, as do the criteria that law firms and legal departments use to determine which model is right for the job.
For some users, the go-to model for a certain legal task might not be the one that performs best, but the one that is best-suited to the task—which are not always the same thing. Instead of gravitating toward a few “winning” generative AI models, it’s likely that the market will turn to a variety of them, including fine-tuned models fit for specific legal purposes.
