AI News

The US Copyright office has solicited comments on ‘‘copyright law and policy issues raised by artificial intelligence systems,’’ drawing thousands of responses. Comments from many authors, and organizations including the Authors Guild and the Association of American publishers, argue that Large Language Models rely on mass copyright infringement in order to function at all. The Authors Guild said,

The leading LLMs today were ‘‘trained’’ on datasets of ap­proximately 200,000 books, plus datasets that include mil­lions of news articles and books scraped from the internet. All the book datasets that we are aware of were copied from pirate ebook sites, and almost all other books that were scraped up by the creators of the datasets compiled from web crawls (e.g., Common Crawl, which has 60 billion domains scraped over 12 years) were also copied from pirate sites.

The AAP concurs, noting that ‘‘several Gen[erative] AI systems have included pirated books in their training datasets’’ concluding that ‘‘fair use should not facilitate wholesale piracy.’’

The responses from tech companies like Google, Meta, and OpenAI, which hope to exploit the technology, defend their methodologies instead. OpenAI says ‘‘the factual metadata and fundamental information that AI models learn from training data are not protected by copyright law,’’ and Meta concurs: ‘‘The extraction of unprotectable facts and ideas from copyrighted works is not itself an infringement of copyright, whether that extraction is accomplished by a human being (by, for example, learning from a book) or by a technological process.’’

Of course, in order to extract that metadata, they mined thousands of illegally sourced copyrighted texts, but OpenAI says they had no choice: ‘‘Technical realities require that copyrighted works be reproduced in order to extract and learn from these unprotectable aspects of a work.’’ OpenAI did make some concessions, explaining that their chatbot ‘‘ChatGPT has been taught in post-training to recognize and decline to respond to certain prompts that appear aimed at reproducing significant portions of works that may be protected by copyright.’’ This ‘‘post-training’’ follows numerous lawsuits from authors who discovered that ChatGPT would, in fact, produce copyrighted material on demand.

These comments and many others can be searched.

One notable response to the Copyright Office was filed by another governmental agency, the Federal Trade Commission, raising numerous issues. The FTC commented further in a press release on the subject. It reads, in part:

The manner in which companies are developing and releasing generative AI tools and other AI products… raises concerns about potential harm to consumers, workers, and small busi­nesses. The FTC has been exploring the risks associated with AI use, including violations of consumers’ privacy, automation of discrimination and bias, and turbocharging of deceptive practices, imposters schemes and other types of scams.

The FTC is concerned ‘‘about the scope of rights and the extent of liabil­ity under the copyright laws. For instance, not only may creators’ ability to compete be unfairly harmed, but consumers may be deceived when authorship does not align with consumer expectations. A consumer may think a work has been created by a particular musician or other artist when it is an AI-created product.’’ The FTC notes that ‘‘conduct that may violate the copyright laws… may also constitute an unfair method of competition or an unfair or deceptive practice, especially when the copyright violation deceives consumers, exploits a creator’s reputation or diminishes the value of her existing or future works, reveals private information, or otherwise causes substantial injury to consumers.’’

The press release notes that ‘‘the FTC has been using its existing legal authorities to take action against illegal practices involving AI, citing con­sumer protection examples including allegations that Amazon and Ring used highly private data they collected to train their algorithms while violating consumer privacy.’’ While they acknowledge AI is still evolving, ‘‘it already has the potential to transform many industries and business practices. No­tably, there is no AI exemption from the laws on the books. Accordingly, the FTC will vigorously use the full range of its authorities to protect Americans from deceptive and unfair conduct and maintain open, fair, and competitive markets.’’ For more.

While you are here, please take a moment to support Locus with a one-time or recurring donation. We rely on reader donations to keep the magazine and site going, and would like to keep the site paywall free, but WE NEED YOUR FINANCIAL SUPPORT to continue quality coverage of the science fiction and fantasy field.

©Locus Magazine. Copyrighted material may not be republished without permission of LSFF.