Before any of this week’s exhibits matter to a judge, somebody has to prove they’re real — and in 2026, that sentence is doing a lot of work. I’m Cassidy, this is Banker Sexual Harassment Lawsuit Watch — and today we’re stepping back from the complaint drama to look at what Federal Rule of Evidence 901 actually asks for before a digital record survives a challenge in a New York courtroom. I’m Brian, and yeah, “901 explainer” sounds dry until you realize this is the part where everything filed this week either holds up or gets shredded. The bar to get something in is low. Keeping it in is where it gets expensive — and that explains a lot about why both sides have been papering the record so aggressively. So if this lawsuit starts pulling in text threads, AI-generated documents, maybe even deepfake allegations, what does a New York court actually want before any of that gets treated like real evidence? Short version: easy to get in, expensive to defend. Under Federal Rule of Evidence 901, a party just has to produce evidence “sufficient to support a finding that the item is what the proponent claims it is” — pretty minimal on paper. But per a Forbes analysis from late April, in 2026 almost anything digital is fair game for a challenge, and the authentication standard that really works — meaning rigorous digital forensics — can be costly and hard to reproduce. Then there’s the separate fight over whether AI-generated documents deserve protection at all: Judge Jed Rakoff of the Southern District of New York ruled in February that documents a defendant created using Anthropic’s Claude and then sent to his lawyer were not protected by attorney-client privilege, because using a commercial AI tool effectively waived that protection. And on the rule-making side, the federal Advisory Committee on Evidence Rules just punted proposed amendments aimed at machine-generated evidence and deepfakes — tabling them until at least fall, per Bloomberg Law, because the committee couldn’t reach consensus. So for now, courts are improvising case by case, often leaning on expert testimony for reliability, with no uniform federal rule in place. So if someone in this JPMorgan case wanted to use AI-generated material — say, a document laying out their version of events — does Rakoff’s ruling mean that could get forced into the open instead of staying confidential? That’s the exposure Rakoff flagged in writing: if a client uses a consumer-facing AI tool and then shares the output with counsel, it likely isn’t privileged, which means opposing counsel could demand it in discovery. Debevoise noted that enterprise AI tools give privilege arguments a somewhat better footing, but even that is no guarantee. So if you’re watching this discovery fight, what each side created digitally — and what tools they used to do it — could end up being as contested as the underlying allegations. If you follow legal fights closely, try Musk v Altman Daily, our daily court-watch on Elon Musk’s trial against Sam Altman, OpenAI, and Microsoft, from testimony to exhibits and the AGI governance fight. Find it wherever you listen to podcasts.
You’ll find links to every story we mentioned today in the show notes. If one stood out, that’s the place to dig a little deeper.
That’s Banker Sexual Harassment Lawsuit Watch for this Tuesday, May 19th. This is a Lantern Podcast.