OpenAI says Apple's sloppy access controls — not theft — explain ex-engineer's messages

OpenAI is publicly pushing back against Apple’s trade secret lawsuit, releasing internal chat logs that paint a very different picture than the one Apple’s lawyers filed in federal court. Rather than a systematic theft of confidential information, the messages suggest a messier reality: Apple’s own employees kept texting their former colleague for help long after he left.

The engineer at the center of the dispute, Chang Liu, left Apple on January 22, 2026. The messages show that on January 27, an Apple employee asked him for a technical assessment, writing, “Of course, I could ask several folks, but you are the best. Even if you don’t work here anymore.” Liu responded with details about internal Apple decisions. Weeks later, on February 14, the same person asked about schematics. In March, Liu found himself added to a group chat with several Apple employees, where he directed them to internal folders and contacts before shutting it down himself: “Hi, this is highly irregular, please remove me from this thread.”

OpenAI’s argument is blunt: Apple’s access management is a mess. They call it a “well-known problem” where “residual access” lingers after employees leave. The blog post also highlights a sloppy lawyering blunder. An outside attorney hired by Apple confused two Asian last names and emailed the wrong person at OpenAI, then falsely claimed to have spoken with OpenAI’s General Counsel by phone. Apple only acknowledged the errors after being corrected, then went silent for five months before suing.

The chat logs are juicy, but they don’t knock down the lawsuit’s biggest pillars. Apple’s complaint names over 400 former employees now at OpenAI and ties the allegations to the startup’s hardware ambitions with io Products, the Jony Ive co-founded venture. OpenAI also rejects claims against Tang Tan, a 24-year Apple veteran, saying he explicitly told his team not to use proprietary information from other companies. The core accusation — that OpenAI actively encouraged departing staff to swipe files — remains untouched by Thursday’s rebuttal. What the messages do show is that the line between poaching talent and protecting secrets is often blurrier in practice than in a legal filing.