Federal judge OKs warrant for ChatGPT records in crypto fraud case; state court separately blocks subpoena

This pair of rulings illustrates the legal patchwork surrounding government and third-party access to AI chatbot records. In the criminal case, a federal judge in the Southern District of New York authorized a warrant for Richard Kim's ChatGPT account, finding probable cause that the records were evidence of securities and wire fraud. The decision opens a new digital surveillance frontier: as AI assistants become integrated into daily life and business communications, the reasoning could expand government access to private user data in ways that chill free expression and erode privacy.

Meanwhile, in the entirely separate civil case Assini v. Hayward, the same Manhattan federal judge blocked a subpoena seeking ChatGPT records, protecting them under the work-product doctrine. The two rulings do not contradict each other—they arise from different legal standards (warrant vs. subpoena, criminal vs. civil) and different courts. But they highlight an urgent need for clear statutory guidelines on when prosecutors, civil litigants, and law enforcement can compel AI platforms to turn over user data, especially as these tools become ubiquitous. Antitrust scholar Tim Wu has warned that concentrated platform power enables extraction of user data and attention; these cases show how that data pipeline can also become a surveillance mechanism.