Sentiment Analysis: Relativity’s Approach to Developing Responsible AI Solutions for e-Discovery and Investigation
As AI technology progresses rapidly, the legal industry faces many ethical questions that will define the years ahead. What AI technology is appropriate for legal use cases? What do consumers need to consider when evaluating AI tools? Who is developing these solutions and what are their objectives in that process?
At Relativity, we recognize the important role we play in developing AI solutions for e-discovery and investigations. This paper outlines how we developed sentiment analysis, a new RelativityOne capability that leverages AI algorithms to detect positive and negative tones and other emotions within data.
Read this white paper to learn how Relativity built a sentiment analysis model that is industry-leading in mitigating bias and principled in our approach to developing AI responsibly. You’ll learn about:
- The importance of sentiment analysis in e-discovery and investigations
- The issues with using historical sentiment analysis models in a legal use case
- Relativity’s approach to developing a “fit-for purpose” sentiment analysis model
- How Relativity’s model compares to other available models in both performance and in bias mitigation