
In that same vein, we expand our discussion to include how Artificial Intelligence tools like Large Language Models (i.e., ChatGPT, Google BARD, Meta’s LLM) could assist legal professionals in their quest to find information, create documents, and help outline legal processes and practices.
He proposed three ways of thinking about the work being done by these models, which are largely analogous to traditional methods. The first way is what Riehl refers to as a “bullshitter,” where a model generates information without providing citations for the information. The second way is called a “searcher,” where a model generates a legal brief, but does not provide citations, forcing the user to search for support. The third way is called a “researcher,” where the model finds relevant cases and statutes, extracts relevant propositions, and crafts a brief based on them.
Riehl believes that option three, being a researcher, is the most likely to win in the future, as it provides “ground truth” from the start. He cites Fastcase’s acquisition of Judicata as an example of how AI can be used to help with research by providing unique identifiers for every proposition and citation, enabling users to evaluate the credibility of the information. In conclusion, Riehl sees a future where AI is used to help researchers by providing a pick list of the most common propositions and citations, which can then be further evaluated by the researcher.
One thing is very clear, we are just at the beginning of a shift in how the legal industry processes information. Riehl’s one-two combination of SALI Standards combined with additional AI and human capabilities will create a divide amongst the bullshitters, the searchers, and the researchers.