{"id":268,"date":"2024-04-24T15:06:51","date_gmt":"2024-04-24T15:06:51","guid":{"rendered":"https:\/\/usatrustedlawyers.com\/blog\/generative-ai-could-revolutionize-e-discovery-but-beware\/"},"modified":"2024-04-24T15:06:51","modified_gmt":"2024-04-24T15:06:51","slug":"generative-ai-could-revolutionize-e-discovery-but-beware","status":"publish","type":"post","link":"https:\/\/usatrustedlawyers.com\/blog\/generative-ai-could-revolutionize-e-discovery-but-beware\/","title":{"rendered":"Generative AI could revolutionize e-discovery, but beware!"},"content":{"rendered":"
\n
\n

<\/p>\n<\/div>\n

Lawyers are typically cautious of new technology and are now diving into \u201ca sea change\u201d of new generative artificial intelligence tools for e-discovery, says Mary Mack, CEO and chief legal technologist at Electronic Discovery Reference Model. \u201cWe are seeing increasing interest in the educational programs we run around generative AI.\u201d<\/p>\n

She adds: \u201cThis change will be rapid, but rapid in legal terms.\u201d<\/p>\n

A large number of vendors have produced generative AI tools for e-discovery. Some, like CoCounsel, offer research and document review within a larger suite of options. LexisNexis has released Lexis+ AI, a case research and document drafting platform. Meanwhile, Lighthouse, Everlaw and Relativity focus specifically on e-discovery tasks.<\/p>\n

Despite a cloud of customer confidentiality concerns, \u201cChatGPT has created this comfort level and belief because [the clients have] played with it, and it creates these magical responses,\u201d says Ian Wilson, founder and CEO of Servient, an e-discovery platform.<\/p>\n

Generative AI’s strengths include sorting documents by talking points and speakers, creating a timeline of who spoke to whom, writing draft reports, and providing links to the actual documents. For example, attorneys using Cloud Court, software that collects and analyzes testimonial outcomes, may learn that \u201cif you ask a question about this specific topic, the witness is likely to pause and become angry,\u201d Mack says.<\/p>\n

\u201cWhether it’s a transcript of a witness statement or a transcript of a hearing, the ability to summarize and use other data sets to say ‘find more like this’ will also have a big impact,\u201d says Janet Sullivan, e-discovery counsel and global advisor to White & Case. director of practical technology.<\/p>\n

Expectations have changed around higher-stakes areas such as privilege logs. A year ago, vendors focused on building tools that helped people make decisions and explain those decisions efficiently, says Jay Worthington, a member of Cloud Court’s board of advisors. \u201cNow they’re all racing as fast as they can to build tools that will automate that decision cycle.\u201d<\/p>\n

Not so fast, says Maura Grossman, a research professor at the computer science school at Canada’s University of Waterloo. Grossman co-authored a groundbreaking article in 2011 on how technology-assisted review, which uses algorithms to classify documents based on input from expert reviewers, could be more effective and efficient in e-discovery than what had previously been done. gold standard was: manual assessment.<\/p>\n

As for generative AI, it remains to be seen \u201cwhether people get a little bored of the hype or whether someone can actually show that it actually has applications that are really efficient and effective,\u201d according to the 2016 ABA Magazine<\/em> says Legal Rebel. “It’s like Star Wars<\/em>, and it looks cool. If it catches on, it’s hard to predict \u2013 or whether consumers will see it as a bigger risk than, say, a tech-enabled assessment.\u201d<\/p>\n

Experts agree that users should stay away from open-source options like ChatGPT. It is known to hallucinate, and it remembers fast and other input information, potentially compromising customer confidentiality.<\/p>\n

Instead, opt for commercially licensed software and enter into contracts with protective orders that stipulate no clues will be stored or information shared, Mack says.<\/p>\n

Craft prompts carefully and don’t ask for information outside the bounded data set, Worthington says.<\/p>\n

\u201cThe risk of ending up in a hallucinatory trap increases exponentially when you step out of bounded data universes,\u201d he adds.<\/p>\n

Legal Rebel Class of 2024<\/h2>\n

Locked up: Criminal justice startups are capitalizing on the early promise of generative AI<\/p>\n

Bridging the gap: Lawyers trying to expand access to justice see promise in generative AI<\/p>\n

Helping hand: Generative AI is already impacting legal research and legal writing<\/p>\n

Always On: Will Generative AI Alleviate Burnout or Make Lawyers Unhappier?<\/p>\n

Head of the Class: Law Schools Consider Post-ChatGPT Courses<\/p>\n

e-Sign on the dotted line: When it comes to using generative AI and contracts, the devil is in the details<\/p>\n

Era of e-Discovery: Generative AI could revolutionize e-discovery, but beware!<\/p>\n

Rewiring Entry: How AI Can Blur the Boundaries of Immigration Law<\/p>\n

This story was originally published in the February-March 2024 issue of the <\/em>ABA Magazine.<\/em><\/p>\n

In this podcast:<\/h4>\n<\/p>\n<\/div>\n