Heppner

Heppner this, Heppner that. I'm just not surprised by the ruling.

It was only a matter of time before some court, somewhere had to decide: is a client's use of Gen or OpenAI covered by the attorney client privilege or the attorney work product doctrine.

On work product: this should be obvious.

A client's own prompts and the resulting answers from AI are not created by counsel--even if done in anticipation of litigation or at counsel's behest, the fact that counsel doesn't do it herself makes this issue clear (at least to me).

On privilege: I also think this should be obvious.

OpenAI is not confidential.

It is textbook legal analysis to say: If you provide otherwise privileged or confidential information to a public, open, not confidential source (or any third-party, without some sort of agreement that applies the privilege beyond you and your client), you waive the privilege.

Maybe the hullabaloo is simply because this is the first written opinion on the matter. I, on the other hand, think this is the obvious outcome.

I think the more interesting question is this:

In civil matters, how many attorneys would think to put something like a client's Claude (or ChatGPT or Grok) conversations on a privilege list to begin with.

That's what triggered the issue in Heppner.

It might be prudent now, for example, to ask all clients--and deponents, and party opponents--

Did you ever talk to any OpenAI source about this matter?

If so, when?

And, what source did you use?

Did you (can you) provide copies of those conversations to counsel?

Did you produce those in this case? If not, we ask that you do so.

Or, to your client: we might have to produce those conversations to the opposing party.

We might be in a new era with AI and all its uses, but these bedrock principles apply and remain unchanged.

✌🏻🔥♥️