跟ChatGPT聊天沒隱私？OpenAI遭控洩露用戶對話給Meta與Google

California woman sues OpenAI for privacy violations, alleging ChatGPT embedded tracking codes that send users’ sensitive conversations to Meta and Google, seeking over $5 million in damages.

A woman in California has filed a class-action lawsuit against OpenAI for violating privacy laws, accusing ChatGPT of transmitting user queries to Meta and Google through tracking tools without user consent. This case is considered the first federal court challenge directly targeting the data handling practices of AI chatbots, sparking widespread concern over AI privacy protection.

ChatGPT suspected of embedding invisible tracking codes, revealing all chat content

On May 13, a California resident representing Amargo Couture filed a lawsuit in the U.S. Southern District Court of California.

The complaint states that OpenAI embedded third-party tracking tools such as Meta’s “Facebook Pixel” and Google’s “Google Analytics” into the code of the ChatGPT official website. These tracking codes operate automatically without user awareness, transmitting users’ query content and personally identifiable information in real-time to Meta and Google, including highly sensitive private data such as health status, financial advice, and legal issues.

Couture stated that she used ChatGPT multiple times between 2025 and 2026 to inquire about personal health and financial matters, unaware that these private conversations had been forwarded to third-party advertising tech giants.

Legal basis: OpenAI’s potential damages exceed $5 million

The complaint alleges that OpenAI violated multiple U.S. and California laws, including the Electronic Communications Privacy Act (ECPA) and California Privacy Act. The scope of CIPA is not limited to traditional telephone communications but also applies to “emerging technologies” such as computers, the internet, and email.

Under California law, the statutory damages for each violation can be up to $5,000. If qualified for class-action status, the potential damages could be substantial, estimated to exceed $5 million. The plaintiffs’ attorneys emphasize that OpenAI assisting Meta and Google in intercepting communications without user authorization constitutes a systemic invasion of privacy for millions of users.

What are tracking codes? Subtle data thieves

“Tracking pixels” are tiny pieces of code embedded in webpage source code that are difficult for ordinary users to detect. Each time a user visits a website, the tracking pixel automatically activates, quietly collecting browsing behavior and operational information, then sending it back to companies like Meta or Google.

For e-commerce platforms or ad-driven websites, this practice is commercially justified; however, the context of ChatGPT is entirely different. Users ask ChatGPT questions out of trust in the platform, often sharing private concerns they wouldn’t tell anyone—such as health issues, emotional pain, or financial crises. If these conversations are recorded by advertising companies, OpenAI risks losing user trust.

Meta leverages this data to build its extensive targeted advertising system. According to the complaint, nearly all of Meta’s revenue comes from advertising, and its ad system can track user activity both on and off the platform, inferring interests, behaviors, and social tendencies to create targeted ad delivery mechanisms.

Urgent need for privacy legislation; AI conversations must not be exploited

As more people incorporate AI tools into daily life, global privacy controversies continue to escalate. Whether it’s users revealing emotional struggles and personal dilemmas to AI chatbots or companies uploading tens of thousands of confidential data points, both individuals and businesses face potential threats.

A few days ago, OpenAI was also sued for allegedly guiding the Florida State University shooter in the ChatGPT platform, indirectly leading to two deaths.

Legal experts believe that if such lawsuits succeed, they will have a significant demonstrative effect on the entire AI industry, forcing companies to more clearly disclose data collection scope, retention periods, and sharing practices, as well as provide explicit consent mechanisms. Currently, OpenAI has not issued any public statement. The case is still in its early stages, and whether it qualifies for class-action status remains to be decided by the court.

• This article is reprinted with permission from 《Chain News》
• Original title: 《ChatGPT Legal Case Adds Another! Accused of Secretly Disclosing User Chats to Meta and Google》
• Original author: Crumax