Apple Sued Over AI Training Data It Says It Doesn't Use

Apple Intelligence and Its Training Data

Apple has been tight-lipped about the specifics of how its AI models are trained, as have most major AI developers. The company launched Apple Intelligence with iOS 18 and has continued expanding the system with each software update. Features include writing tools, image generation, photo cleanup, and integration with ChatGPT for more complex requests.

Apple has said that its on-device models are trained using a combination of licensed data and synthetic data generated by Apple itself — an approach designed to limit legal exposure and improve privacy. The company has not disclosed a comprehensive list of data sources, which is standard across the industry.

The dataset referenced in the Chicken Soup for the Soul lawsuit appears to be a separate, open-web corpus that circulates widely in AI research. Apple's position is that even if data from that corpus exists somewhere in its research infrastructure, it did not flow into the models that constitute Apple Intelligence as a product.

Broader Legal Landscape

The AI copyright litigation wave shows no signs of slowing. Courts are still working through foundational questions — including whether training an AI on copyrighted text constitutes fair use, what damages might look like at scale, and whether outputs from AI models constitute derivative works.

The outcomes of landmark cases against OpenAI and others are being watched closely by the entire industry. A ruling that finds AI training categorically infringing would have sweeping consequences. A ruling that upholds a broad fair use defense would, conversely, settle the legal environment in favor of AI developers for years to come.

For Apple, the stakes extend beyond this single lawsuit. The company has positioned privacy and responsible AI development as core brand values. Being associated with large-scale unauthorized data use — even if the allegation is ultimately dismissed — runs counter to that narrative.

What Comes Next

The case is in early stages, and no court date has been set. Apple is expected to challenge the lawsuit on the grounds that the specific dataset cited was not used in its AI products, potentially seeking early dismissal.

Legal analysts note that Apple's proactive disclosure — publicly stating the dataset doesn't power Apple Intelligence before the suit gained traction — is a strategic move to set the record straight early. It also signals that Apple anticipated this line of legal attack and prepared a response in advance.

Whether that defense holds will depend on what discovery reveals about Apple's actual training pipeline. If Apple's internal data sourcing records support its public claims, the publisher's path to damages becomes narrow. If they don't, the company faces the same legal exposure that has ensnared other AI developers.

The lawsuit underscores a growing tension in the technology industry: AI systems require enormous quantities of text to reach commercial viability, and the legal framework for acquiring that text responsibly is still being written in real time.

This article is based on reporting by 9to5Mac. Read the original article.