Topic: Ai Copyright

Mustafa Suleiman, CEO of Microsoft AI, claimed that content on the open web has been "freeware" since the 90s and is fair use for AI training. This assertion is challenged by the hosts, who cite established copyright law and personal legal victories against unauthorized commercial use of digital content.

ByteDance's new AI video model, Seedance 2.0, has sparked controversy by creating realistic deepfake fight scenes featuring Tom Cruise and Brad Pitt. SAG-AFTRA President Sean Astin condemned the technology for infringing on the likenesses and voices of actors without authorization. The union is pushing for the "No Fakes Act" in the Senate to protect performers' rights against AI-generated replacements.

Paul McCartney and over 1,000 other musicians released a protest album titled "Is This What We Want" to oppose AI-related copyright legislation in the UK. The digital release consists of silent studio recordings, while the vinyl edition features ambient noises like tape hiss and footsteps. The project serves as a warning that the creative ecosystem will collapse if AI companies exploit intellectual property without compensation.

AI startup Anthropic has agreed to a $1.5 billion settlement in a class-action lawsuit brought by authors who alleged the company used pirated books to train its models. While the settlement is large, a San Francisco judge's ruling that training AI constitutes "fair use" is seen as a significant legal victory for the industry.

A federal judge dismissed a copyright lawsuit brought by authors including Sarah Silverman and Ta-Nehisi Coates against Meta. The judge ruled that the plaintiffs failed to prove that Meta's Llama AI model violated the law by training on their works. The ruling noted that the authors' lawyers made the "wrong arguments" regarding market flooding and copycat content.

Disney and Universal have filed a landmark copyright infringement lawsuit against the AI imagery company Midjourney. The studios allege that the AI was trained on their iconic characters, such as those from *Cars* and *Toy Story*, without authorization. A key legal question is whether the liability rests with the AI company or the individual users who generate the prompts to create infringing content.

NPR reported on the abrupt firing of the Librarian of Congress and the head of the U.S. Copyright Office following the release of a report on AI and fair use. The report suggests that using copyrighted works to train AI may qualify as fair use on a case-by-case basis. The hosts discuss the unreliability of AI tools like Perplexity and ChatGPT, noting they often provide incorrect information depending on how a question is phrased.

Brewster Kahle of the Internet Archive discussed the potential for AI to make digital libraries more interactive and searchable. Despite ongoing copyright lawsuits in the U.S., the organization continues to archive over a billion URLs daily, with the Wayback Machine now hosting 900 billion pages. The archive also preserves unique collections, such as 78rpm records and the recently deleted MTV News digital archives.

New legislation has been proposed to require AI developers to disclose the sources of their training data to protect copyright holders. While some argue that scanning all human knowledge into a database is a societal benefit, others worry about the loss of intellectual property rights. Additionally, industry insiders claim that AI-generated code is often inferior to that produced by human programmers.

Artists are using a new tool called "Nightshade" to sabotage AI training models by "poisoning" image data. The tool alters pixels in a way that is invisible to humans but causes AI to misinterpret prompts, such as returning an image of an egg when a user requests a red balloon.

Signal President Meredith Whittaker criticized the narrative surrounding generative AI, describing it as a hype cycle for cloud services. She noted that AI models often produce "janky" text with no relationship to facts or citations and generate images that infringe on artists' work. Whittaker argues that while AI can replace junior copywriters, it remains unreliable for serious professional contexts.

The music industry is reacting to viral AI-generated tracks that mimic the voices of artists like Rihanna and Drake. Universal Music Group has requested that streaming services block AI companies from scraping their catalog to train models. The hosts argue that modern pop music is already so processed that it is easily replicated by algorithms.

A host tests the beta version of Google Bard, which provided inaccurate personal information about Adam Curry's family and history of being banned from Twitter. The discussion covers the high cost of running AI queries and the looming legal battles regarding plagiarism and copyright violations by AI models.