CRIV CONNECTION

The CRIV semiannual call with Thomson Reuters took place Wednesday, December 13, 2023. This update includes news, product enhancements, and Q&A. 

Participants: Mary Jenkins (CRIV Liaison to Thomson Reuters), Kim Hurley (Thomson Reuters Information Management Consultant), Dave Corbett (Westlaw Product Development), Valerie McConnell, (Thomson Reuters Senior Director, Client Success), Zach Ratzman (Director, Practical Law), Vani Ungapen (AALL Executive Director), and Michelle Hook Dewey (AALL Executive Board Liaison to CRIV) 

AI Assisted Research on Westlaw Precision

Dave Corbett described how AI-Assisted Research harnesses the power of generative AI in Westlaw Precision, emphasizing security commitments and continuing product developments. He noted that Thomson Reuters’ privacy and security policies are aligned with ISO and NIST standards, and that Thomson Reuters expressly prohibits any vendor from retaining or using Westlaw customer data to train their generative AI models. The company is taking a risk-based approach to development and assessment and is committed to transparency and to pre-processing and post-processing steps that reduce hallucinations and improve accuracy. Dave commented that Westlaw Precision’s AI-Assisted Research is not intended to generate work product; it’s about providing responsive information and authoritative support to significantly accelerate the research process. For example, citations in the answers not only link to the supporting authority, but directly to the language that produced the answers as well as materials for further exploration; it is not simply listing authorities. It accommodates follow-up questions as well as starting new research, allows delivery via download, print, and email, and makes previous queries and their answers available via History. Information is readily available about Westlaw’s AI-Assisted Research, along with tips for best results.

Brief Overview of CoCounsel Skills

Valerie McConnell demonstrated CoCounsel Core, an AI platform customized for legal practice, built by Casetext using GPT-4 and subjected to rigorous testing. It is trained for law, summarizes documents, and has guardrails for liability. She assured participants that CoCounsel is private and secure, with data storage and processing on dedicated servers. CoCounsel elements include (1) general skills—timelines, summaries, draft review; (2) litigation skills—prepare for a deposition, correspondence, and responding to discovery; and (3) transactional skills —extraction of contract data, policy compliance, and more. It addresses many use cases including docket review, document preparation, contract review, and evaluation of potential liability. The administrator can create a list of linked databases for search purposes. Valerie demonstrated CoCounsel’s deep review and analysis using prompts, phrases, and natural language, showing its application for redlining and reviewing, for example, third party documents like contracts, and its production of timelines, reports with governing law and policies, and recommendations for bringing a document into compliance.

Practical Law Updates on Artificial Intelligence

Zach Ratzman highlighted Practical Law’s generative AI capabilities—Ask Practical Law AI—which uses a conversational chat experience to accelerate research. This new functionality, launching in January 2024, allows users to ask legal questions and quickly receive succinct, synthesized answers that are created exclusively using Practical Law content. Those answers contain footnotes to the underlying Practical Law resources, which allows users to easily verify the answer, while also providing a springboard into additional relevant Practical Law content.  Zach also previewed Practical Law’s Artificial Intelligence Toolkit, which published earlier in 2023 and aggregates content to help legal professionals better handle legal questions flowing from the use of AI.  The Toolkit includes links to numerous Practical Law resources on AI, including several trackers, which track legal developments relating to AI, such as litigation, legislative, and regulatory proposals and changes; a model employment policy on the use of generative AI in the workplace; and other up-to-date, expertly-written resources on AI.

QUESTIONS & ANSWERS

Q. We understand that academic institutions’ plans are not included in the Thomson Reuters generative AI rollout at this point. Please explain plans or decisions for Thomson Reuters’ generative AI exposure or access in the academic sector. Which user groups or library types have it or will have access over the next six months (e.g., will the judicial and government offices where law students are clerking or externing have access to Thomson Reuters’ generative AI tools?)

A. The head of the Westlaw academic group notes that generative AI capabilities are to be announced and rolled out in the second half of 2024.

Q. How is Thomson Reuters supporting its customers administratively in demonstrating value and ROI via quantitative analysis of a platform’s use or underutilized resources? Does Thomson Reuters see customer access to detailed usage information as a value add for the company and for the customer? Why or why not?

A. Thomson Reuters is committed to providing administrators with usage data via QuickView and Westlaw Analytics, TR’s billing tools. QuickView displays both in-plan and out-of-plan usage. One can target QuickView reports to generate information on a particular content set. Both tools demonstrate a commitment to transparency, granularity, and detail. Westlaw Analytics is often used to conduct detailed billing investigations. In the academic context, one can couple the reporting with a graphical history to export and share.

Q. Recently, some organizations were notified of the elimination of one publisher’s content or publications from Westlaw. We understand that it is a contractual matter between vendors; this is not an inquiry about the specifics of that situation. Is dropped content by customer type or sector typical and what generally are the reasons for content changes based on customer type? What are the general policies on notifying subscribers about content removal?

A. Thomson Reuters does not single out particular segments when dropping content from Westlaw. This has never been a practice we follow. Aspen Publishers, Wolters Kluwer content, was removed from Westlaw due to low content usage across all Westlaw subscribers. When content becomes redundant or ceases to provide value to our customers, we make adjustments to our offerings. Similar content remains available on Westlaw, and we are continually evaluating and investing in new titles to ensure that our customers have the best content that meets their organization’s research needs.

Q. What are the circumstances under which a library maintenance agreement like West Complete might be modified and in what ways, apart from financial exigency? With reductions in law libraries’ purchase of print, does Thomson Reuters see value in more flexibility around print subscriptions, like subbing out content of equal value or other approaches in order to retain that business?

A. Thomson Reuters values our long-standing relationship with your organization and we appreciate an opportunity to respond to your question. We strive to offer innovative contract terms that best support our customers’ evolving needs. Our agreements balance considerations related to flexibility, budgetary limitations, and market dynamics. We continually review and evaluate contracting options that best fit the current market conditions. Our LMA contracts do offer a “swap” period annually where customers may replace existing subscriptions with new/different titles that better fit their needs. Customers are encouraged to reach out to their client executives or regional managers to discuss LMA terms.

CRIV appreciates Kim Hurley’s coordination of this conversation on Thomson Reuter’s end.