Latest news with #copyright
Wall Street Journal
37 minutes ago
- General
- Wall Street Journal
News Quiz for June 28, 2025
This copy is for your personal, non-commercial use only. Distribution and use of this material are governed by our Subscriber Agreement and by copyright law. For non-personal use or to order multiple copies, please contact Dow Jones Reprints at 1-800-843-0008 or visit
Times
8 hours ago
- Politics
- Times
Danish citizens to ‘own their own faces' to prevent deepfakes
Denmark plans to become the first country in the world to give its citizens copyright over their faces and voices in an effort to clamp down on 'deepfakes' — videos, audio clips and images that are digitally doctored to spread false information. In recent years the tools for making deepfakes, including artificial intelligence-assisted editing software, have become so sophisticated and ubiquitous that it takes not much more than a few clicks of a mouse to create them. They are already endemic in the political sphere and were deployed during recent election campaigns in Slovakia, Turkey, Bangladesh, Pakistan and Argentina. The former US president Joe Biden was subjected to an audio deepfake during the Democratic presidential primary in New Hampshire last year. In November an MP from the German Social Democratic party was reprimanded for posting a deepfake video of Friedrich Merz, the conservative leader and future chancellor, saying that his party 'despised' the electorate. The Danish culture ministry said it would soon no longer be possible to distinguish between real and deepfake material. That in turn would undermine trust in authentic pictures and videos, it warned. 'Since images and videos swiftly become embedded in people's subconscious, digitally manipulated versions of an image or video can establish fundamental doubts and perhaps even a completely wrong perception of genuine depictions of reality.' There is now broad cross-party support in Denmark's parliament for a reform to the copyright law that would make it illegal to share deepfakes. The bill includes a special protection for musicians and performing artists against digital imitations. 'We are now sending an unequivocal signal to all citizens that you have the right to your own body, your own voice and your own facial features,' said Jakob Engel-Schmidt, the culture minister. Lars Christian Lilleholt, the parliamentary leader of the Danish Liberal party, which is part of the ruling coalition, said AI tools had made it alarmingly easy to impersonate politicians and celebrities and to exploit their aura of credibility to propagate false claims. 'It is not just harmful to the individual who has their identity stolen,' he said. 'It is harmful to democracy as a whole when we cannot trust what we see.' The reform will include an exemption for parody and satire. This is a thorny area: several studies suggest a large proportion of political deepfakes are humorous or harmless rather than malicious. There are some experts who warn that concern about the phenomenon risks tipping over into a moral panic. In April last year Mette Frederiksen, Denmark's Social Democratic prime minister, was targeted with an AI-generated deepfake that fell into this grey area. After her government announced that it was abolishing a Christian public holiday, the right-wing populist Danish People's Party released a video of a fake press conference where Frederiksen appeared to say she would scrap all the other religious holidays, including Easter and Christmas. The clip, which was presented as a dream sequence and clearly labelled as AI-manipulated content, prompted debate about the acceptable boundaries of the technology.
Yahoo
12 hours ago
- Business
- Yahoo
Federal judge denies OpenAI bid to keep deleting data amid newspaper copyright lawsuit
NEW YORK — A federal judge has upheld a ruling directing OpenAI to preserve logs and data slated for deletion after news outlets including the Daily News suing the technology giant accused the company of hiding evidence of copyright infringement. The new ruling, issued Thursday in Manhattan Federal Court, denied the company's objection to an earlier court order directing OpenAI to keep any data used to train its artificial intelligence bots — logs which plaintiffs say may contain details of widespread content piracy. OpenAI executives have maintained that they are merely safeguarding users' privacy by objecting to any data retention request or order. But lawyers for the plaintiffs said the privacy argument is nothing more than a distraction. 'This is like a magician trying to misdirect the public's attention,' said Steven Lieberman, a lawyer representing the News and several other media outlets. 'That is absolutely false. The judge has made clear and plaintiffs have made clear that they don't want to receive information that personally identifies the users of these conversations. If data is turned over, it will only be turned over anonymously. And OpenAI knows that. No one's privacy it's at risk.' The publishers' key argument at the core of their lawsuit is that the data that powers the company's popular ChatGPT has included millions of copyrighted works from the news organizations. The publications have argued that such content has been used without consent or payment — which translates to copyright infringement on a massive scale. Various reports have placed the company's value at $300 billion. making it one of the most valuable private companies in the world, thanks in part to its online chatbox, ChatGPT, which was released in 2022. But when it comes to raw material — redistributed creative content — OpenAI took the cheap and easy way out, Lieberman said. 'They just stole it from the newspapers, from magazines and from book authors,' he said. A representative from OpenAI did not immediately respond to a request for comment. OpenAI has argued that the vast amount of data used to train its artificial intelligence bots is protected by 'fair use' rules. The doctrine applies to rules that allow some to use copyrighted work for purposes like criticism, commentary, news reporting, teaching and research. However, lawyers for the newspapers have argued that the fair use test involves transforming a copyrighted work into something new, and the new work cannot compete with the original in the same marketplace. The court has rejected OpenAI's position that the newspapers haven't produced 'a shred of evidence' that people are using ChatGPT or OpenAI's API products to get news instead of paying for it. The New York Times originally brought the suit in December 2023. The News, along with other newspapers in affiliated companies MediaNews Group and Tribune Publishing, filed in April 2024. The other outlets included The Mercury News, The Denver Post, The Orange County Register and the St. Paul Pioneer Press, and Tribune Publishing's Chicago Tribune, Orlando Sentinel and South Florida Sun Sentinel.
TechCrunch
18 hours ago
- Business
- TechCrunch
Big Tech lands an early win in legal battles against publishers
This week, two major AI companies scored early wins in court, with federal judges siding with Meta and Anthropic in separate lawsuits over how their models were trained on copyrighted material. The decisions represent the first real legal validation of AI companies' argument that training models on books, images, and other creative works can be considered 'fair use' — even if those materials weren't obtained with permission. It's a big deal for companies building generative AI, and a potential turning point for the many lawsuits still in motion. Today, on TechCrunch's Equity podcast, hosts Max Zeff and Anthony Ha were joined by Sean O'Kane (who graciously stepped in while Kirsten headed off to the Nevada desert to see the next big act of Redwood Materials, the startup founded by former Tesla CTO JB Straubel) to dive deeper into the rulings. While neither case sets a precedent yet, Anthony noted that appeals are likely, and broader challenges could ultimately shape how AI companies interact with entire industries going forward. Listen to the full episode to hear more highlights from the week, including: Kalshi's $185M round, and what it says about the rising (and legally murky) world of prediction markets The startup betting on reusable satellites, and why the Department of Defense is paying attention Tesla's robotaxi rollout in Austin, and how it stacks up against Waymo and other AV companies' approaches Equity will be back next week, so stay tuned! Equity is TechCrunch's flagship podcast, produced by Theresa Loconsolo, and posts every Wednesday and Friday. Subscribe to us on Apple Podcasts, Overcast, Spotify and all the casts. You also can follow Equity on X and Threads, at @EquityPod.
The Verge
a day ago
- Entertainment
- The Verge
What Meta and Anthropic really won in court
A lot of the future of AI will be settled in court. From publishers to authors to artists to Hollywood conglomerates, the creative industry is picking a big copyright fight over the vast quantities of data used to train AI models — and the ultimate output of those models. (Disclosure: Vox Media, The Verge's parent company, has a technology and content deal with OpenAI.) This week, we got rulings in two early cases, involving groups of authors suing Anthropic and Meta. In both cases, the tech companies won. Sort of. On this episode of The Vergecast, Nilay, David, and Jake talk a lot about the twin rulings this week, and whether the AI companies may have won the battle without winning the war. But before we get to all that, there's some other tech news to talk about! We run through the first few days of the Tesla robotaxi rollout and the latest on the Trump Phone, both of which are going about as you'd expect. We talk about the new Fairphone 6 and Titan 2, two fascinating but maybe slightly niche ideas about smartphones. And we talk about Meta's new face computers, one made with Xbox and one made with Oakley. Subscribe: Spotify | Apple Podcasts | Overcast | Pocket Casts | More After that, The Verge's Adi Robertson joins the show to dig into the AI cases. We talk through the ways the plaintiffs failed to make the right arguments, and why the judges in both cases appear desperate for someone to come in and do better. We talk about the difference between buying books and pirating them, between inputs and outputs, and the actual creative risks that come from flooding the internet with AI slop. Finally, in the lightning round, it's time for another round of Brendan Carr is a Dummy, some debatably huge news about HDMI, and the end of the Blue Screen of Death. The blue-ness of it, at least. If you want to know more about everything we discuss in this episode, here are some links to get you started, beginning with the gadgets of the week: And in AI lawsuit news: And in the lightning round:
