AI Chats May Be Used Against You: Lawyers Warn

Lawyers warn AI chats may be used as legal evidence.

Legal Cautions Regarding AI Chat Privacy

Legal Cautions Regarding AI Chat Privacy US lawyers warn that your chats could be used against you after an AI verdict. April 15 (Reuters) - Some American attorneys are advising their clients not to use AI chatbots like trusted confidants when their freedom or legal culpability is at stake, as more and more people look to AI for guidance.After a federal judge in New York decided this year that the former CEO of an insolvent financial services company could not protect his AI chats from prosecutors filing securities fraud charges against him, these warnings became even more urgent. Following the decision, lawyers have advised that prosecutors in criminal cases or litigation opponents in civil cases may compel discussions with chatbots such as Claude from Anthropic and ChatGPT from OpenAI.We are informing our customers: Alexandria Gutiérrez Swette, a lawyer with the New York-based legal firm Kobre & Kim, advised proceeding cautiously in this situation. Under U.S. law, conversations between individuals and their attorneys are nearly always considered secret. However, since AI chatbots are not attorneys, lawyers are advising their clients to take precautions to protect the privacy of their interactions with AI tools.

Court Decision and Legal Consequences

Court Decision and Legal Consequences A JUDICIAL DECISION Bradley Heppner, the founder of alternative asset business Beneficent (BENF.O) and former head of bankrupt financial services company GWG Holdings, was involved in the case that helped raise red flags. Heppner entered a not guilty plea after being accused of securities and wire fraud by federal authorities in November of last year. Heppner prepared reports about his case using Anthropic's chatbot Claude, which he then shared with his lawyers. The lawyers later contended that Heppner's AI exchanges should be kept since they included information regarding his defense.Because Heppner's defense attorneys were not directly involved and because attorney-client privilege does not apply to chatbots, the prosecution contended that they had the right to seek anything that Heppner developed with Claude. The conventional legal protections for attorney communications may be compromised if a lawyer voluntarily discloses information to any other party. In February, Manhattan-based U.S. District Judge Jed Rakoff ordered Heppner to turn over 31 case-related papers created by Anthropic's chatbot Claude. There is no legal relationship "or could exist, between an AI user and a platform such as Claude," according to Rakoff.

AI in Legal Cases and Divergent Court Opinions

AI in Legal Cases and Divergent Court Opinions The increasing use of artificial intelligence by attorneys and self-represented parties in court is already causing courts to struggle. Among other things, this has resulted in court filings that contain fictitious cases created by AI. Rakoff's ruling was a crucial early test for fundamental legal safeguards governing interactions between attorneys and their clients as well as litigation-related information in the AI chatbot era. On the same day as Rakoff's decision, U.S. Magistrate Judge Anthony Patti in Michigan ruled that a woman defending herself in a lawsuit against her former employer was exempt from having to turn up her conversations with OpenAI's ChatGPT on the employment allegations raised in the case. Instead of treating the woman's AI chats as discussions with someone her company may try to utilize for its defense, Patti treated them as a component of her own "work-product" for the case. In his order, Patti stated that ChatGPT and other generative AI programs "are tools, not persons."

Advice from Attorneys and Potential Consequences

Advice from Attorneys and Potential Consequences Attorneys compete to establish guards. Both OpenAI and Anthropic's usage and privacy policies allow them to share user data with outside parties. Additionally, both say that before using their chatbots for legal guidance, consumers must speak with a trained expert. During a hearing in Heppner's case in February, Rakoff observed that Claude "expressly provided that users have no expectation of privacy in their inputs." Lawyers' guidance has ranged from advising clients to carefully choose their AI platforms to recommending certain wording for chatbot prompts. In client advisory, Los Angeles-based O'Melveny & Myers and other companies stated that "closed" AI systems intended for business usage could offer more robust security for legal communications, albeit even that is yet mostly untested. According to some firms, conducting AI legal research under a lawyer's supervision increases the likelihood that it will be protected by attorney-client privilege. A person should indicate in the chatbot prompt if a lawyer does suggest using AI. Until then, lawyers claim that a long-standing presumption still holds true: Avoid discussing your matter with anyone other than your attorney, including AI.