Law Business Research’s Post

Definitely things to start thinking about if you're not already - thanks Chuck Kellner for this article. I've been stewing about the proposed FRE 901(c) since reading this yesterday and would love to discuss with everyone and learn more. Just a paralegal's two cents here, but I feel that draft language of the rule is just a hair narrower gateway rather than a barrier when a benchmark is what is needed. If a court would be convinced that the evidence is more than likely fabricated/altered in whole or in part, the proffering party shouldn't necessarily be then given courtesy of the court's consideration as to whether or not its probative value outweighs its prejudicial effect. Could the evidence be denied admission altogether unless any portion of the evidence can be rehabilitated through an independent expert? Or through some other foundational, non-interested non-party? Fabricating or altering evidence is intended generally to deceive (pull the smoking gun out of thin air, or hide it?) or distract (add one red herring or two?). The "evidence" was more than likely skewed in the highly probative direction from the start, no?

🚨 In-house lawyers will not need to outsource the examination of every piece of evidence for #deepfakes. There are ways they can filter evidence themselves. Read more in this Legal Dive article by Everlaw Strategic Discovery Advisor, Chuck Kellner! 👇 👇 👇 #Ediscovery #LegalTech

What are your thoughts on using #AI in the courtroom? While it has the potential to speed up trial prep and automate systems, nothing beats the human touch. Let's ensure AI complements, not replaces, our skill sets. Read more from CNBC. ⬇

Attorneys know Practical Law as the trusted, up-to-date legal know-how platform that helps them get accurate answers. Its enhancement, Ask Practical Law AI is a generative AI search tool that dramatically improves how you access trusted expertise. https://ow.ly/XvzN50Qu6l6

The Judicial Conference’s Advisory Committee on Evidence Rules has released its report on artificial intelligence, exploring the potential need for modifications to the Federal Rules of Evidence. While the committee didn't propose new rules, it identified key areas for further consideration, including reliability of machine learning output, authentication of deepfakes, addressing deepfake claims, and validation studies for machine learning evidence. https://lnkd.in/gNBf--bN #AI #machinelearning #lindastanley #evidence #legaltechnology #deepfakes #legalnews #justicesystem

Stanford's Human-Centered AI group lampooned Thomson Reuters and LexisNexis genAI tools because of their susceptibility to hallucinations. This is one of the main solutions JuristAI's software suites solve. https://lnkd.in/evbyDCA7 Hallucination-Free? Assessing the Reliability of Leading AI Legal Research Tools https://lnkd.in/e7fN4W-S

Here are 5 ways the new generation of Lexis+ AI will help legal professionals transform their work, from faster results to and even better user experience: https://lnkd.in/gBA9cmvS

Troy In The Press: Check out Business Insider's coverage of our Legix Beta! Learn more about our proprietary tech; what makes us different is what makes us better. https://lnkd.in/gxB85ky4 #ai #AIInnovation

AI in 2024: What Every GC Needs to Know

New Guidance from the USPTO on the use of #AI in proceedings before it. As in previous memos, there is a heavy focus on the duty of candor and good faith for any party that submits papers thereto. I am happy to see the USPTO invoke the #signature requirements. Every #patent practitoner should know (but many do not!) that no one can sign your name for you. Not a secretary, administrative assistant, or an AI. It is improper. The signature certifies that to the best of the signer's knowledge the information therein is factually and legally true. AI #LLMs are prone to hallucination and making things up. You see the point. The Guidance warns that if AI is used to draft anything "material to patentability," that would need to be disclosed. Particularly if it drafts claim elements or embodiments other than the inventors' contributed. The Guidance also notes that using AI to autopopulate IDS could be a filing for an improper purpose if the signer did not review the submission. "First, 37 CFR 1.4(d) requires a natural  person to personally sign or insert their signature on the IDS. By signing, that person is certifying that they have performed a reasonable inquiry—including not just reviewing the IDS form but reviewing each piece of prior art listed on the form ... After the contents have been reviewed,  clearly irrelevant and marginally pertinent cumulative information to the instant proceeding  should be removed to avoid violating 37 CFR 11.18 by overburdening the examiner with a large  amount of irrelevant information." The Guidance also highlights the duty to keep client information confidential. A company's promise that it will may not be enough. Per the guidance "AI systems may retain the information that is entered by users. This information can be  used in a variety of ways by the owner of the AI system including using the data to further train  its AI models or providing the data to third parties in breach of practitioners’ confidentiality obligations to their clients under, inter alia, 37 CFR 11.106. If confidential information is used to train AI, that confidential information or some parts of it may filter into outputs from the AI system provided to others." It also notes AI servers outside the US could trigger foreign filing license and export requirements. Of course this is true outside the context of AI.