Neurocientistas processam Apple por uso supostamente ilegal de livros piratas no treinamento de IA

Gigante da tecnologia enfrenta ação judicial de pesquisadores cerebrais

Acusação: Apple utilizou material protegido sem autorização para alimentar seus sistemas de inteligência artificial

Os neurocientistas alegam que a empresa contornou direitos autorais ao incorporar milhares de obras acadêmicas em seus algoritmos

A briga judicial expõe as práticas questionáveis por trás da corrida corporativa pela supremacia da IA

Mais uma demonstração de que, no mundo corporativo, a ética frequentemente perde para o lucro - como sempre acontece quando há dinheiro envolvido

Apple’s AI push and the ‘shadow library’ problem

Apple Intelligence is Apple’s suite of AI-powered features built into its devices, such as the iPhones, iPads, and Macs. The complaint notes that the day after Apple officially announced Apple Intelligence, its market value rose by more than $200 billion, “the single most lucrative day in the history of the company,” the lawsuit said.

The plaintiffs allege that Apple, like other AI developers, uses shadow libraries, which are massive online repositories that store plagiarized and unauthorized copies of academic, scientific, and literary works. The claimants say that their work was used without permission.

Apple has not publicly disclosed the full range of datasets used to train Apple Intelligence. The lawsuit argues that this lack of transparency masks the company’s reliance on illegally sourced data, in contrast to its marketing emphasis on user privacy and responsible innovation.

AI firms face a wave of copyright challenges

In early September, a group of authors filed a copyright suit against Apple, claiming that their works were used without consent to train Apple’s AI tools. The company joins a growing list of technology firms, including OpenAI, Microsoft, Meta Platforms, and Anthropic, that have faced lawsuits over the unauthorized use of copyrighted content to develop generative AI systems.

As Cryptopolitan reported, Anthropic itself agreed in September to pay $1.5 billion to settle a separate class action lawsuit brought by authors over the training of its chatbot Claude.

AI developers claim that the process they use to train their models qualifies as fair use because it transforms the data to learn linguistic or visual patterns rather than reproducing the original works. 

However, authors and publishers have a different view of the matter, as they maintain that the copying involved is extensive and commercial in nature, violating their rights and undermining their creative livelihoods.

High stakes for Apple and the AI industry

For Apple, which entered the generative AI race later than its rivals, the lawsuit threatens its effort to position Apple Intelligence as a privacy-focused, ethically trained alternative to other chatbots, especially when the accusations are in stark contrast to what it projects itself to be.

Legal experts say the outcome of these cases could change how AI systems are built and trained. A ruling against Apple or similar firms could compel companies to license copyrighted works at scale, significantly increasing development costs and potentially slowing innovation. 

On the other hand, a ruling favoring tech firms could entrench the fair-use precedent and limit compensation options for creators.

Want your project in front of crypto’s top minds? Feature it in our next industry report, where data meets impact.