Apple has filed a new legal complaint in the Northern District of California alleging that former engineer Chang Liu used an autonomous AI agent to analyze stolen circuit schematics. The company claims this use of artificial intelligence could result in permanent and unquantifiable damage to its proprietary hardware designs.
The autonomous AI agent running LTspice simulations
The use of an autonomous AI agent to process stolen data represents a new and potentially permanent threat to corporate intellectual property. According to the new court filing, Chang Liu allegedly used a company-issued MacBook to download highly confidential Apple circuit schematics and subsequently uploaded them into LTspice, a software program used for simulating circuit performance before physical manufacturing begins.
Apple’s legal team argues that the involvement of an AI agent significantly aggravates the alleged misconduct.. A text message reportedly sent by Liu in March suggests he deployed an autonomous agent to run these simulations on LTspice without direct human oversight. This distinction is critical because Apple contends that feeding trade secrets into an AI model that learns from the input could create a "continually propagating" use of the intellectual property, making the damage effectively irreversible and impossible to quantify.
Alleged forensic destruction talks with Yu-Ting Peng
Apple is seeking to expand its investigation into Chang Liu's conduct following allegations of evidence tampering. The legal complaint states that in June, Liu allegedly discussed the destruction of forensic data with his OpenAI colleague, Yu-Ting Peng. This discussion reportedly occurred after Liu became aware that he was a subject of Apple's scrutiny.
To address these concerns, Apple's attorneys are now requesting the expedited surrender of a Mac Mini.. The company believes this specific hardware was used to power the LTspice simulations and is necessary to determine the full extent of the alleged misconduct. The filing emphasizes that the full scope of the breach remains unknown, necessitating a deeper forensic dive into the device.
OpenAI's defense citing Apple's AI talent struggles
OpenAI has pushed back against these accusations, characterizing the lawsuit as a meritless attempt to stifle competition. In a response filed last month, OpenAI suggested that Apple's legal actions are motivated by the tech giant's difficulty in competing for AI talent and its perceived failure to integrate advanced artificial intelligence into its own consumer products.
The legal battle also involves Tang Tan, another former Apple employee who moved to OpenAI last summer.. Apple accuses Tan of encouraging other departing staff to discuss confidential information during the hiring process. However, OpenAI has countered that Apple has failed to identify any specific trade secrets that were actually communicated during these interviews, arguing that Tan could not have violated the federal Defend Trade Secrets Act without a specific, identified secret being shared.
The unverified scope of the Mac Mini data
Significant questions remain regarding the exact nature of the information processed by the AI agent. While Apple has identified the use of LTspice, it has not yet publicly detailed which specific circuit schematics were compromised or how much of that data was ingested by OpenAI's broader machine learning models.
Furthermore, the full extent of the alleged coordination between Chang Liu and Yu-Ting Peng remains unverified. It is currently unknown whether the discussions regarding forensic data destruction were successful or if other employees were involved in the alleged attempt to obstruct Apple's investigation. The outcome of this case may set a massive precedent for how autonomous systems handle proprietary corporate data.
Comments 0