Hurwitz Fine Buffalo and Long Island Litigation Attorneys Michael Williams and Patrice Melville recently presented a webinar exploring the opportunities—and risks—of using artificial intelligence in the legal profession. Below are the key takeaways:
1. AI is a tool, not a substitute for professional judgment.
Generative AI can provide a valuable starting point for research and analysis, but attorneys remain responsible for evaluating the facts, applying the law, and determining whether the information produced is accurate.
2. Every AI-generated output must be independently verified.
AI can fabricate cases, quotations, facts and legal propositions. Any information generated by an AI platform should be cross-checked against reliable sources before it is relied upon or incorporated into legal work.
3. Closed AI systems are not risk-free.
While closed systems may offer greater control over access, data retention and security, both open and closed AI platforms can produce inaccurate, incomplete, biased or fabricated information.
4. Protect confidentiality when using AI.
Attorneys should understand where information goes after it is entered into an AI platform and should never assume that information submitted to a system will remain private or confidential.
5. Data minimization is the safest approach.
Avoid entering entire documents or identifying information when it is unnecessary. Use only the information needed for the task, anonymize facts where possible, and remember that removing a client's name may not be enough if other details could identify the client or matter.
6. Existing professional responsibilities still apply when using AI.
The convenience of AI does not replace an attorney's duties of confidentiality, competence, supervision, candor and independent professional judgment.
7. Attorneys remain responsible for everything they submit.
Regardless of the tools used to prepare a document, lawyers must review and verify the facts and law before presenting it to a court, opposing counsel or another party, and be sufficiently attentive to identify hallucinated law or facts in opposing documents to better assist the court.
8. AI can reproduce bias contained in its underlying data.
AI-generated responses may reflect historical assumptions and biases found in the information on which the technology relies, making independent evaluation of its output essential.
9. AI errors can have serious professional consequences.
Courts have addressed filings containing nonexistent cases and other AI-generated inaccuracies, reinforcing the importance of cite-checking, due diligence and candor when errors occur.
10. “Actual intelligence” remains essential.
AI can increase efficiency and help lawyers process significant amounts of information, but the attorney must ultimately review the work, exercise judgment and make informed decisions based on verified facts and law.
