In United States v. Heppner, Judge Rakoff of the Southern District of New York ruled that written exchanges between a criminal defendant and generative AI platform Claude were not protected by attorney-client privilege or the work product doctrine.
That's one judge. An audio tape made by a criminal defendant is intended for review by his counsel is a non-discoverable privileged communication. The tape retains this character if reviewed by an attorney-authorized paralegal. What difference exists where the attorney has the tape summarized by AI. I respectfully submit that Hizzoner is incorrect.
We might also ask if the best venue to decide national AI regulation is a single judge sitting in a criminal case involving a fraudster. If Judge Rakoff is correct, then a trade secret shared with AI is no longer a trade secret. This affects not just a single NY criminal defendant, but anyone that runs a company and wants to keep business practices secret. I would submit that this is no way to regulate a field such as AI.
I dunno, ruling seems to have a point to me, a non-lawyer. claude is not an attorney and his attorney was not involved until after the claude "conversation". Look at for instance the exchange of emails with your lawyer, are they privileged? Yes, with caveats according to gemini but it could be lying of course. How about if you emailed your mobster uncle asking for advice on how to use a lawyer to keep your guilty ass outta jail? Is that privileged? All of a sudden I'm not so sure.
This seems to be a pretty narrow ruling but maybe I'm missing something not being a lawyer and all.
It seems the key here isn't—or shouldn't be—what kind of service the defendant used, but whether something special happens when a service is involved in preparing a message to his lawyer.
IMO if the "for my lawyer" purpose/intent is not in dispute, then it shouldn't matter whether the service is a search-engine, an LLM, a browser-based word processor, or the drafts/sent folders of a webmail client.
The reverse direction is much clearer: Imagine a client receives an obviously-privileged email from their lawyer, and uses a cloud text-to-speech service to listen to it. Should that audio/text be admissible as evidence? Hell no.
>whether something special happens when a service is involved in preparing a message to his lawyer.
I use an online LLM to field better questions to my lawyer — he is aware, as I send him these AI conversations. His only warning to me is don't say anything that you wouldn't want the judge to read, which is the same warning given about email. Lot's of "devil's advocating" phrasing...
During our current lawsuit (my first, as plaintiff) — years brewing — I have built myself a local Ollama computer, which can answer offline questions better. But for something quick or simpler, I still use online services often.
----
Local LLMs are the future. This massive datacenter overinvestment is going to become obvious (similar to how ProTools destroyed recording studios).
This is a really interesting and well written case update/critique. I agree with the author's that the judge's reliance on Anthropic's fine-print privacy policy does not satisfy the actual legal standard governing privilege. Or if it did, it would raise extremely thorny issues around all of the cloud-based technology products that lawyers and clients use every day.
That said, I note that the court's opinion specifically calls out Anthropic's practice of *training models on user data* as a reason why the defendant could not have expected confidentiality. I do not use these cloud models for anything important precisely because they are operated by companies, like Anthropic, that are completely untrustworthy.
7 comments
[ 8.7 ms ] story [ 32.9 ms ] threadWe might also ask if the best venue to decide national AI regulation is a single judge sitting in a criminal case involving a fraudster. If Judge Rakoff is correct, then a trade secret shared with AI is no longer a trade secret. This affects not just a single NY criminal defendant, but anyone that runs a company and wants to keep business practices secret. I would submit that this is no way to regulate a field such as AI.
This seems to be a pretty narrow ruling but maybe I'm missing something not being a lawyer and all.
IMO if the "for my lawyer" purpose/intent is not in dispute, then it shouldn't matter whether the service is a search-engine, an LLM, a browser-based word processor, or the drafts/sent folders of a webmail client.
The reverse direction is much clearer: Imagine a client receives an obviously-privileged email from their lawyer, and uses a cloud text-to-speech service to listen to it. Should that audio/text be admissible as evidence? Hell no.
I use an online LLM to field better questions to my lawyer — he is aware, as I send him these AI conversations. His only warning to me is don't say anything that you wouldn't want the judge to read, which is the same warning given about email. Lot's of "devil's advocating" phrasing...
During our current lawsuit (my first, as plaintiff) — years brewing — I have built myself a local Ollama computer, which can answer offline questions better. But for something quick or simpler, I still use online services often.
----
Local LLMs are the future. This massive datacenter overinvestment is going to become obvious (similar to how ProTools destroyed recording studios).
That said, I note that the court's opinion specifically calls out Anthropic's practice of *training models on user data* as a reason why the defendant could not have expected confidentiality. I do not use these cloud models for anything important precisely because they are operated by companies, like Anthropic, that are completely untrustworthy.