Extract from HaystackID’s article “Lessons from the OpenAI Litigation on Safeguarding Privileged Information”
In the age of Artificial Intelligence, lawyers are struggling to apply traditional procedural and evidentiary rules to AI-related evidence. Disputes over discovery involving such information are proliferating, with the In re OpenAI, Inc., Copyright Infringement Litigation lawsuit being the epicenter for these disputes. While OpenAI recently resolved various issues regarding the preservation and production of user prompts and AI outputs, the court just issued another order addressing the application of the attorney-client privilege to prompts and outputs.[1] That order, which rejects a claim of privilege over such information, highlights steps litigants may consider to better safeguard privileged information from disclosure in litigation. OpenAI is also instructive on whether technical issues relating to the training of AI models, along with communications exchanged with in-house lawyers over collaboration platforms like Slack, can be claimed as privileged.
The Attorney-Client Privilege and Privilege Logs
By way of background, the attorney-client privilege is a rule of evidence that typically prevents communications between clients and lawyers from being disclosed in court proceedings.[2] For a document to be privileged, it must be confidential, generally limited to the client and lawyer, and made for the purpose of seeking or providing a legal opinion or advice.
While the attorney-client privilege is a rule of evidence, procedural rules govern how parties claim the privilege in civil litigation.
