SBIR Phase II: Artificial Intelligence Tool for Analysis of Legal Documents

The broader impact of this Small Business Innovation Research (SBIR) Phase II project will be to reduce the need for attorney-driven expertise and an extended discovery period, and instead shift reliance to data-driven expertise to enhance and democratize outcomes in civil cases. The proposed AI platform will allow for the computational prediction of case outcomes, helping to better inform case decisions and address existing inefficiencies within the legal industry. Adoption of this technology will allow attorneys to bypass the time-intensive process involved in case value prediction, allowing for more time to focus on discovery and strategy, as well as allow for better informed decision-making. Notably, the platform will have built-in anti-bias algorithms that will actively work to correct discriminatory past outcomes when making data-driven computations moving forward. Therefore, this technology is not only the first to unlock access to the wealth of informative yet disparate data to support personal injury attorneys in quickly making reliable decisions but is also alone in delivering anti-bias tools that improve fairness and access to justice for clients from all backgrounds, most notably among demographics that have not historically received equal or fair compensation. This Small Business Innovation Research (SBIR) Phase II project aims to improve the likelihood of a positive outcome for attorneys’ clients, while also freeing up time for attorneys to take on greater caseloads, resulting in social justice gains. In virtually every business sector, data analysis has become a driving force behind decision-making, yet the legal services sector has largely lagged, with many law firms instead relying on conventional wisdom and time-intensive research. To address this issue, the preceding Phase I project leveraged an innovative learning technique to enable algorithm training across multiple decentralized databases without exchanging data samples, thus keeping information private and confidential. This Phase II project seeks to 1) Build out the artificial intelligence (AI) model to return predictions on value and outcomes over case lifetime and improve accuracy of trial predictions; 2) Expand platform features to support case management and improve usability; and 3) Build a novel dataset compiling demographic information on 50,000 case outcomes to quantify the racial component of case bias and develop bias-correcting models. This project will significantly expand the applicability of the AI-driven platform to deliver reliable, equitable, and interpretable outputs on case value prediction for diverse case types. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.