Raising Suspicion: The Telltale Signs a Party Has Used AI

Law practitioners are seeing a growing trend of encountering correspondence, such as demand letters and court filings, that appear to be AI-generated, raising concerns about quality and authenticity. While some firms address these suspicions strategically, others prioritize responding to substantive claims over pointing out technical flaws. The use of generative AI in legal work remains a topic of debate, with some practitioners emphasizing its potential when carefully refined, and critics warning against reliance on unpolished outputs.

Sterne Kessler Director Dan Block says he’s encountered demand letters that appear to be AI-generated, based on the language used and letter structure. “But it’s hard to confirm that generative AI is being used,” he says.

He says he’s more interested in whether parties are using generative AI to do infringement analysis than whether they’re using it to help draft letters. “But in the cases where I feel like it’s been used, [the letters are] generally lower quality,” he adds.

Block continued to say that he would highlight if someone was citing improper case law in a demand letter. However, he would be unlikely to point out typos. “I don’t think that helps move the ball forward unless the typos would change the claim,” he says. That said, he wouldn’t necessarily respond to errors differently based on whether he suspected the other party of using AI or if he assumed human error had occurred.