Apple Sues OpenAI Over an Alleged Insider Bug That Kept a Fired Engineer's Access Alive for Weeks
A software glitch let a departed Apple engineer keep reading confidential company files after he joined OpenAI. Apple says that was no accident.
Key points
- Apple filed a lawsuit on Friday seeking court orders to stop OpenAI from using confidential information it says was stolen by former Apple employees.
- A software bug, described by Apple as "rare", temporarily allowed a terminated employee to retain access to Apple's internal servers for weeks after leaving the company.
- The lawsuit names former Apple engineer Chang Liu, who Apple says spent eight years on some of its most sensitive product programs before joining OpenAI.
- Apple alleges OpenAI conspired with former Apple staff to take what it calls an "unlawful shortcut" toward building AI-powered consumer devices.
- A second Apple employee, Yu-Ting "Alyssa" Peng, is named in connection with internal messages exchanged with Liu during the alleged scheme.
Apple filed a lawsuit Friday accusing OpenAI of conspiring with former Apple employees to steal trade secrets, trade secrets being confidential business information that a company is legally entitled to protect. The case, first reported by Ars Technica, centres on an unusual technical failure that Apple says made the alleged theft possible.
When an employee leaves a company, their access to internal systems is normally cut off immediately. Here, Apple says a rare bug in its own systems failed to do that. The departing engineer, Chang Liu, could still read files on Apple's servers for weeks after his termination, even after he had already started working at OpenAI.
Liu is not a peripheral figure. Apple says he spent eight years working on some of its most sensitive product development programmes before moving to OpenAI.
Apple discovered the problem while investigating internal messages between Liu and a then-current Apple employee named Yu-Ting Peng, also known as Alyssa Peng. Apple does not say in its complaint exactly what those messages contained, but the company used them as a starting point for a wider investigation.
The goal, Apple alleges, was competitive sabotage at scale. In its complaint, Apple says OpenAI and the former employees worked together to "take an unlawful shortcut" toward launching a line of AI-powered consumer devices as commercially attractive as the iPhone. That is a significant claim: it positions the alleged theft not as opportunistic, but as a deliberate strategy.
OpenAI has not yet filed a public response to the lawsuit.
Should ordinary people be worried about this?
This case does not directly threaten consumers, but it shows how insider access, even access that should have ended, can sit open far longer than anyone realises. Bugs in offboarding systems are more common than companies like to admit.
If you work somewhere with access to sensitive data, ask your IT team whether departing colleagues lose access on their last day or whether that process is manual and occasionally missed. If you manage a team, review your own offboarding checklist now.
The lawsuit asks the court to block OpenAI from using any information allegedly taken through this route. How much of OpenAI's current product development that might touch is, at this stage, unknown.
Watch for: companies announcing policy changes around employee offboarding and access management in the coming months, particularly in the technology sector, as legal scrutiny on this issue grows.
