On July 22, John Tredennick, Merlin Search Technologies founder and CEO, presented a one and one half hour program at the International Arbitration Club of New York. The program focused on the transformative potential of large language models, using ChatGPT as the example, to streamline investigations and ediscovery.

The program highlighted salient questions surrounding the use of large language models in investigations and ediscovery. He also demonstrated real-world game-changing ways in which ChatGPT will be used to move the industry forward from traditional search to true discovery—finding answers in large document collections.

The program answered the following questions:

• What can ChatGPT do? And what it can’t do.
• What are the limits? Should we be worried about privilege or confidentiality?
• How can you use prompts? Are your prompts training GPT?
• What practical, real-life ways can GPT and other large language models streamline ediscovery and investigations?
• Can you point GPT at your document collection and turn it loose, like a typical TAR algorithm?
• Will ChatGPT replace human reviewers? 

Download a PDF of the slides John presented for answers to these questions and additional insights into this emerging technology. You’ll also get a sneak peek into what the Merlin lab is doing with ChatGPT. 

John walks through real-world experiments and results that he and Merlin Chief Data Scientist Dr. William conducted using just ChatGPT and a basic search engine. Experiments include building searches, analyzing and summarizing documents and spreadsheets, creating investigation reports, summarizing depositions, and even analyzing chat messages. 

Stay tuned for more exciting things to come.