logo
#

Latest news with #EuropeanArtificialIntelligenceAct

EU AI Act doesn't do enough to protect artists' copyright, groups say
EU AI Act doesn't do enough to protect artists' copyright, groups say

Euronews

time02-08-2025

  • Business
  • Euronews

EU AI Act doesn't do enough to protect artists' copyright, groups say

As the European Artificial Intelligence Act (AI Act)comes into force, groups representing artists say there are still many loopholes that need to be fixed for them to thrive in a creative world increasingly dominated by AI. The AI Act, celebrated for being the first comprehensive legislation to regulate AI globally, is riddled with problems, these organisations say. Groups like the European Composer and Songwriter Alliance (ECSA) and the European Grouping of Societies of Authors and Composers (GESAC) argue that it fails to protect creators whose works are used to train generative AI models. Without a clear way to opt out or get paid when tech companies use their music, books, movies, and other art to train their AI models, experts say that their work is continually at risk. 'The work of our members should not be used without transparency, consent, and remuneration, and we see that the implementation of the AI Act does not give us,' Marc du Moulin, ECSA's secretary general, told Euronews Next. 'Putting the cart before the horse' The purpose of the AI Act is to make sure AI stays 'safe, transparent, traceable, non-discriminatory and environmentally friendly,' the European Commission, the European Union's executive body, says in an explainer on the law. The law rates AI companies based on four levels of risk: minimal, limited, high, or unacceptable. Those in the unacceptable range are already banned, for example AIs that are manipulative or that conduct social scoring, where they rank individuals based on behaviour or economic status. Most generative AI falls into a minimal risk category, the Commission says. The owners of those technologies still have some requirements, like publishing summaries of the copyrighted data that companies used to train their AIs. Under the EU's copyright laws, companies are allowed to use copyrighted materials for text and data mining, like they do in AI training, unless a creator has 'reserved their rights,' Du Moulin said. Du Moulin said it's unclear how an artist can go about opting out of their work being shared with AI companies. 'This whole conversation is putting the cart before the horse. You don't know how to opt out, but your work is already being used,' he said. The EU's AI Code of Practice on General-Purpose (GPAI), a voluntary agreement for AI companies, asks providers to commit to a copyright policy, put in place safeguards to avoid any infringements of rights, and designate a place to receive and process complaints. Signatories so far include major tech and AI companies such as Amazon, Google, Microsoft, and OpenAI. AI providers have to respect copyright laws, the Commission says The additional transparency requirements under the AI Act give artists clarity on who has already used their material and when, du Moulin added, making it difficult to claim any payment for work that's already been scraped to train AI models. 'Even if the AI Act has some good legal implications, it only works for the future – it will not be retroactive,' Du Moulin said. 'So everything which has been scraped already … it's a free lunch for generative AI providers who did not pay anything'. Adriana Moscono, GESAC's general manager, said some of her members tried opting out by sending letters and emails to individual AI companies to get a license for their content, but were not successful. 'There was no answer,' Moscono told Euronews Next. 'There was absolute denial of the recognition of … the need to respect copyright and to get a license. So please, European Commission, encourage licensing'. Thomas Regnier, a Commission spokesperson, said in a statement to Euronews Next that AI providers have to respect the rights holders when they carry out text and data mining, and if there have been infringements, they can settle it privately. The AI Act 'in no way affects existing EU copyright laws,' Regnier continued. Mandate licence negotiations, groups ask Du Moulin and Moscono are asking the Commission to urgently clarify the rules around opting out and copyright protection in the law. 'The code of practice, the template and the guidelines, they don't provide us any capacity to improve our situation,' Moscono said. 'They're not guaranteeing … a proper application of the AI Act'. The advocates said the Commission could also mandate that AI companies negotiate blanket or collective licenses with the respective artist groups. Germany's Society for Musical Performing and Mechanical Reproduction Rights (GEMA) filed two copyright lawsuits against AI companies OpenAI, the parent of ChatGPT, and Suno AI, an AI music generation app. While not directly related to the AI Act, Du Moulin says the verdict could determine to what extent AI companies could be bound to copyright laws. The Commission and the European Court of Justice, the EU's high court, have also signalled that they will review the text and data mining exemption in the copyright legislation issued in 2019, Du Moulin said. New AI companies have to make sure they are compliant with the AI Act's regulations by 2026. That deadline extends to 2027 to companies already operating in the EU.

Spain Tests the Waters on Artificial Intelligence
Spain Tests the Waters on Artificial Intelligence

Yahoo

time16-05-2025

  • Entertainment
  • Yahoo

Spain Tests the Waters on Artificial Intelligence

In March, Spain's government became one of the first countries in Europe to approve a draft law concerning AI, almost a year to the date after formal approval of the landmark European Artificial Intelligence Act provided a common legal framework for the development, commercialization and use of AI systems across Europe. AI is one of the most divisive issues in the entertainment industry today. A week after Spain's draft law was unveiled, 400 Hollywood creatives signed a letter of concern about copyright protections for the arts and entertainment sector, pushing back against OpenAI and Google's appeals to the U.S. government to allow their AI models to train on copyrighted works. More from The Hollywood Reporter Ari Aster's 'Eddington' Gets Muted Response at Starry Cannes Debut with Joaquin Phoenix, Emma Stone and Pedro Pascal Ready for a Change of Scenery? Try Côte d'Azur Town Surfing 'Arco' Review: In Charming, Natalie Portman-Produced French Animated Film, a Boy From the Future Must Find His Way Home Meanwhile, James Cameron recently suggested filmmakers could save 50 percent on big-budget films by using AI, to which Netflix CEO Ted Sarandos replied he hoped AI could also make them '10 percent better.' In a smaller industry like Spain's, these are powerful arguments. 'Technological advances are a welcome and important boost for an industry accustomed to fighting against giants, like high production budgets and excessive bureaucracy,' says Beatriz Pérez de Vargas, director of AI Alter Ego, the Invisible Intelligence, a three-chapter docuseries on AI for public broadcaster RTVE, which also won awards for its use of AI. Daniel H. Torrado, director of The Great Reset, which he bills as Spain's, if not Europe's, first entirely AI-generated feature, agrees. 'All creators have a ton of projects in the closet, because many of them either can't find funding or there are production issues to get them off the ground,' he says. His apocalyptic tech thriller — about an AI borne from the mind of a rogue hacker that threatens to destroy the world — 'would have had unaffordable costs and production times' without AI, he says. 'AI allowed us to simulate complex decisions early on and experiment without the budgetary risk that often paralyzes many independent creators.' But, Torrado adds, 'human oversight was constant. Every artistic, narrative andemotional decision went through my hands. AI was a powerful tool, not a substitute for the creator.' That's a theme among those experimenting with AI right now. 'We need to embrace it, but it cannot replace art,' says film and commercials director Paco Torres, who gives AI training sessions to private companies and government organizations around the world. 'We cannot lose artists, the white paper, the creation from nothing, the emotions, the human interactions, the imperfection … We need to fail, to not be perfect — this is important because it's how we get emotion.' Regulation in Spain Underscoring the catch-22 of regulating AI, José Enrique Lozano, creator of a new AI and big data master's program at Madrid's School of Cinematography and Audiovisual (ECAM) and managing director of AV data and consulting firm GECA, says striking the right balance will be challenging: 'If we want to protect ourselves from artificial intelligence to maintain our status quo … Spain and Europe need to be more aggressive and make much more progress in regulating artificial intelligence. On the other hand, I think the more we regulate, the more we'll fall behind.' Manuel Cristóbal, director of the Seville European Film Festival and a longtime producer specializing in animation, agrees. 'We have to see AI as an opportunity,' he says. 'If you create restricted laws on different continents and in different countries, that would kill creativity … and it will be developed elsewhere.' Namely, in places with less regulation, which is why a global conversation about where we see ourselves as a species is warranted, Torres suggests. 'Do we have a philosophy set up to say, OK, this is happening to us, and how are we going to take this?' The RTVE series AI Alter Ego, the Invisible Intelligence features experts from around the world examining AI from two perspectives: utopian and dystopian. The show won two Lovie awards (a pan-European prize recognizing excellence in internet culture) for its use of 10 different tools to create 2D and 3D visuals as well as to write lines of script and create AI characters. The series' director, Pérez de Vargas, says there is little consensus about how to move forward. 'In general, the European perspective reflects a much more preventive approach,' she observes. 'Other countries, such as the United States and China, are more pro-innovative. Their priority is not to slow technological development.' Lily Li, a lawyer at California-based Metaverse Law specializing in privacy and AI, notes that 'in Spain and in other countries, you're seeing draft AI bills that align to the EU AI Act. There is more of a push toward harmonization. In the United States, we are definitely seeing fragmentation between different states regarding their approach to AI. We do not have federal AI legislation or a cohesive approach to AI at the federal level.' Li suggests the U.S. would also do well to follow the EU Act's inclusion of AI literacy. Among the greatest risks of AI, according to experts and the people interviewed for this story, are questionable veracity (deepfakes and misinformation), subliminal manipulation of human behavior, lack of diverse opinions, invasion of privacy, lack of transparency and concentration of power. Li points specifically to 'high-risk AI processing' impacting individuals' well-being, such as decisions about their employment, insurance or credit. The new Spanish draft law for the 'ethical, inclusive and beneficial use of artificial intelligence' addresses these concerns, mirroring the EU Act language and setting up a series of steep fines, ranging from $7 million to $35 million, or between 2 percent and 7 percent of a firm's annual global turnover (less for small businesses), for what it considers harmful practices. It's set to be overseen partly by the Spanish Agency for the Supervision of Artificial Intelligence, formed in 2023. 'AI is a very powerful tool that can be used to improve our lives or to attack democracy; it can have good or bad usages,' Óscar López, Spain's minister for digital transformation and public service, said at the March 11 news conference introducing the law, calling the proper use and governance of AI 'crucial.' Steps are also being taken elsewhere in the local entertainment industry. The Spanish Film Institute is adding language barring projects using generative AI tools from vying for some of its subsidies. In June, Spain's Film Academy voted to ban AI-generated soundtracks from competing for the country's top Goya Awards. AI Productions Torrado presented his AI feature, The Great Reset, at the European Film Market in Berlin, and he'll show it to potential distributors at the Cannes Marché as well. 'We integrated artificial intelligence tools into virtually every phase: script, conceptual design, pre-visualization, visual planning, image generation, postproduction and editing,' he says. 'We used a combination of custom generative models to maintain stylistic coherence.' The workflow, he says, was 'more similar to an animated production than a traditional shoot, but it shortens timelines, reduces costs and enhances narrative precision.' The film cost less than $230,000, largely spent on AI subscriptions and copyrights, compared with the $8 million in Spain or maybe $50 million it might have cost in the U.S., he says. Actors were hired for one week to use as 'references' for movement or voice. From start to finish, with an already completed script, the English-language film took about six months to make. 'I think AI can amplify creativity,' Torrado says. 'AI opens up dozens of possible routes. It's like having a huge team of creative assistants at your disposal who never rest and always have new ideas. The challenge is knowing how to filter, decide, curate and guide.' He adds that he's been experimenting with AI for the past two and half years, working with such programs as ChatGPT, Midjourney, Runway, ElevenLabs, Pixabay, Krea and more on short films and ad and spec campaigns. Yet he says he takes a 'protectionist' stance, learning and using the tools as 'an extension of my creativity for how I'm going to shoot,' but setting limits and only using AI in preproduction. He calls it a 'hybrid' model and insists on always having a team of humans in the mix. 'Authorship remains in the hands of the creator,' he insists. 'AI has neither originality nor creativity; in fact, automation is the opposite of creativity and originality. It's repetition, patterns, repeating processes … At the end of the day, AI is just that: a tool for a creator.' Training the Next Generation Werner Herzog once famously proclaimed, 'If you want to do a film, steal a camera, steal raw stock, sneak into a lab and do it!' Cristóbal cites the quote: 'Not anymore. We all have a camera right in our pockets, but not everybody does films or narrative films that are worth watching. AI will be the same. It will be a tool.' He envisions a new generation of filmmakers arising with these tools. In Spain, training options for that new generation are beginning to pop up. The Universitat Autònoma de Barcelona offers a bachelor's degree in AI. RTVE offers a specialized course in AI as part of its training institute, focused on creation as well as literacy. Madrid film school ECAM launched its new AI and big data program in September. 'The master's degree was conceived to provide students with the practical skills necessary to address the challenges of the digital revolution in the audiovisual environment, specifically in the area of ​​audience data analysis, consumer habits and metrics,' says Alejandra Álvarez Suárez, ECAM's head of continuing education and postgraduate studies. Adds program director Lozano: 'In general in the industry, we need to incorporate people who know how to speak both languages, who are bilingual in that sense — capable of speaking the language of both hemispheres, that of data science and that of understanding the content and the people, who see that content from a more human, more social point of view.' Finding the Boundaries Cristóbal calls himself 'an early adopter' as executive producer on the first CGI film in Europe (2001's The Living Forest) and the first day-and-date release in Spain (2006's Going Nuts). 'I have always been involved in technology and storytelling,' he says. While not fearful about AI, he adds, 'I think it's a whole new world, and of course we need to put some boundaries to it.' He cites deepfakes and copyright — AI's use of existing material — as key concerns. 'If somebody is using AI to write a script, that's their choice. [But] if there's a copyright infringement, they will have to deal with it.' Torrado agrees, adding that the need to establish guardrails is vital. 'I'm also concerned about the lack of transparency in model training. We've used tools and services with clear licenses and professional ethics, but not all companies can say the same. Some models are trained with copyrighted material without consent or compensation, and that puts creators at risk. It's urgent to establish an ethical and legal framework that protects both art and artists.' He and Li both also point to some tools, like Sora and DeepSeek, which are not available everywhere in Europe because of legal or bureaucratic obstacles. 'This puts us at a disadvantage compared to other, more agile industries,' Torrado says. For Pérez de Vargas, the streaming platforms' use of predictive audience and trend analysis to decide what to produce worries her. 'This could lead to a decline in innovative projects,' she says. 'Producers' intuition will become increasingly less relevant, and this will lead to audiovisual products becoming increasingly similar and homogeneous. Along the way, we will miss out on serendipities that could spearhead new voices and different ways of storytelling, which will no longer have a place in our sector.' And of course, there's no turning back — the genie is very much out of the bottle. 'Artificial intelligence cannot be 'uninvented,' ' Torrado says. 'It is here to stay, and refusing to use it is like giving up electricity or the internet. It's a transformative tool that, if well managed, can democratize cinema and open doors to new voices. But to achieve this, we need courage, knowledge, intelligent regulation and, above all, the political and cultural will to embrace innovation.' As one expert interviewed in the debut episode of the RTVE series puts it, 'Whoever has this technology more developed will probably be the one who writes history and the future.' Best of The Hollywood Reporter 'The Goonies' Cast, Then and Now "A Nutless Monkey Could Do Your Job": From Abusive to Angst-Ridden, 16 Memorable Studio Exec Portrayals in Film and TV The 10 Best Baseball Movies of All Time, Ranked

DOWNLOAD THE APP

Get Started Now: Download the App

Ready to dive into a world of global content with local flavor? Download Daily8 app today from your preferred app store and start exploring.
app-storeplay-store