Apple sees a breach of its confidential technology and a threat to its future hardware work; OpenAI portrays the case as an overreach built on Apple’s own muddled access controls. The dispute now extends beyond the courtroom, with each side pressing a sharply different account of what former employees did—and why.

The fight began with Apple’s lawsuit last month against former iPhone engineer Chang Liu, former Apple design executive Tang Tan, and OpenAI. Apple alleges that Liu and Tan took secret information about unreleased technologies, processes and products to support OpenAI’s hardware ambitions. Its accusations include claims that Liu retained a company computer, accessed Apple storage after leaving, downloaded confidential files and advised a colleague how to avoid drawing the attention of Apple’s security team.¹

On Monday, Apple escalated by seeking a preliminary injunction that would bar Liu, Tan and OpenAI from accessing, acquiring, using or disclosing the disputed material while the litigation proceeds.¹ Apple also alleges that Tan sought confidential details during interviews with Apple employees being recruited for OpenAI.

OpenAI answered not with a court filing, but with a public post and selected iMessages and emails. It called the lawsuit “careless, aggressive, and oddly personal,” arguing that Apple had failed to shut off former employees’ access and then tried to characterize the resulting problem as misconduct.²

The company’s central rebuttal is blunt: “Apple’s request for a preliminary injunction is both based on false information and completely unnecessary because we do not have, nor want, any of their trade secrets.”² It says Tan made clear that OpenAI “do[es] not want, and must not use, any confidential information from other companies.”²

OpenAI further says Apple’s lawyers initially contacted the wrong person after confusing two Asian surnames, undermining Apple’s suggestion that OpenAI ignored outreach. Apple, meanwhile, is asking the court to treat the alleged files and recruitment conduct as an immediate risk—not merely a communications failure. The next phase will test whether Apple’s evidence outweighs OpenAI’s public receipts.