AI is increasingly a dominant focus in the legal technology conversation. As evidence of this, recent legal technology forums turned swiftly to discussions about how generative AI tools could be applied within the field of electronic discovery (eDiscovery). Highlighting this trend, the announcement of aiR for Review by Relativity during their keynote at the latest RelativityFest conference made headlines.
Relativity’s Chief Product Officer, Chris Brown, shared the company’s plans to bring the most recent developments in Generative Pre-training Transformer (GPT) technology to document reviews. It is already being used by a select number of clients for testing, with a wider release planned for the end of the year.
According to Above the Law, aiR for Review promises to significantly speed up stages of litigation or investigative review. It is expected to deliver detailed case analysis with both greater speed and quality. Users will be able to locate data responsive to a discovery request, identify materials relating to different legal issues and key documents, and view the rationale behind aiR for Review’s recommendations.
The function of aiR for Review appears impressively user-friendly. Users simply direct the system to review for responsiveness, provide a synopsis of the issue at hand, input some crucial terms, and grant it an idea of what denotes a hot document. Once these details are provided, the tool can autonomously carry out its work. Highlighted passages along with rationales are then made available for the user to verify. Additionally, the user can enhance results by adding commentary, which the tool learns and applies in refining the search, thereby creating a constant feedback loop.
Reviewers remain the brains of the operation, and aiR offers transparency by allowing users to see exactly what the AI sees. This promotes an iterative process of refining AI’s approach to the data set. Chris Brown noted that this narrow use of technology helps avoid common dangers such as misalignment or hallucinations that can occur with artificial intelligence.
As the AI technology evolves, the ability of artificial intelligence to autonomously and effectively review documents may become a reality, albeit with a need for careful deliberation around its ethical implications and safe employment.