The EU AI Act is a changemaker, Jimmy Stewart and Maria Callas get the AI treatment, and other news
The EU AI Act Aims to be a Global Regularity Benchmark
The European Union has provisionally agreed on its landmark AI regulation. The legislation incorporates safeguards for general-use AI, limitations on biometric identification systems, bans on manipulative AI and social scoring, and allows the right for consumers to complain. There are hefty fines dished out for non-compliance to big-tech and banned applications, including biometric categorization systems, that use sensitive personal data and likenesses. European Commission President Ursula von der Leyen has called the Act the first comprehensive global legal framework on AI. The Act aims to anticipate new rules, invite big tech companies to join the voluntary AI Pact, and prioritize safety, fundamental rights, and support human-centric, responsible, transparent AI development. The new AI legislation will provide a global benchmark that safeguards the benefits and potential risks of AI. The agreement has to pass formal adoption by Parliament and Council to become EU law, and this can take 1–2 years.
Calm Sleep App Features ‘It’s a Wonderful Sleep Story’ with an AI Jimmy Stewart
Frank Capra’s It’s a Wonderful Life starring Jimmy Stewart as an American everyman who dreams of far-off places until business and family obligations push him to end his life. Stewart’s performance is one of the reasons the film still resonates 77 years after its first release to mediocre reviews. Surely AI couldn’t replace the actor, who died in 1997? The sleep app Calm has taken a stab at it with permission from Stewart’s estate and recreated the actor’s voice for a new sleep story. Kyiv-based startup, Respeecher, who created Mark Hamill’s voice for The Mandalorian, created Stewart’s voice with the help of a living actor. CEO Alex Serdiuk wanted to use a living actor to adhere to “high ethical standards.” Okay, but AI or a living impersonator is still not the real Jimmy Stewart.
Meta Goes Public with AI Image Generator Trained on Instagram and Facebook Photos
Meta’s text-to-image AI generator was released earlier this year on its Instagram and Facebook platforms. This latest public release is a standalone service outside of its messaging platform called Imagine. Meta’s AI model called Emu, is trained on 1.1 billion Instagram and Facebook user photos. Yes, your photos. The controversy around copyright violation and data privacy that has rocked leading tech and image platforms has forced Meta to state that private messages and images that were not shared publicly, were excluded from the data crawl. Time will tell if Meta will come under fire again for using data without user permission.
Maria Callas XR/AI Recreation Marks 100th Anniversary
Maria Callas’s likeness has been recreated with AI technology as part of an ongoing XReco European project. Vienna-based production company FFP also recreated the interior of the Budapest Opera House where Hollywood actor, Angelina Jolie recently filmed scenes for a new film Maria about the opera singer’s life. Her synthetic image was reconstructed from 50+ internet images, generative AI, AI-based shape estimation methods, traditional view-based 3D modeling, and a trained Stable Diffusion model to pose Callas from any angle. Callas was one of the biggest performers of the 20th century and the project aims to push new groundbreaking technology while introducing Callas to a contemporary audience.
“We are entering a new era, where 3D modeling for XR, film and special effects will become obsolete, being replaced by Neural Rendering and Diffusion Models,” Barnabas Takacs, Ph.D. from FFP.
Maria Callas — watch the video
CBC News Report Uncovers Thousands of Canadian Book Authors Used in AI Training Dataset
Some of the biggest authors, such as Margaret Atwood, Alice Munro, and Gordon Korman, are among 2,500 copyrighted books written by 1,200 Canadian authors whose work has been shared online by a now-defunct dataset used for AI training and research. None of the authors were asked for permission to use their work. This follows an American investigation of some of their big-name authors who were involved in a class action with the company responsible, Books3. George R.R. Martin, John Grisham, and Jodi Picoult were among the action. The biggest pain point is that their authored content could be used to generate new content in the style of a distinctive author’s voice. Osgoode Hall Law School professor Carys Craig isn’t convinced that Books3 is breaking any copyright law because AI datasets are mainly used to understand patterns in language, which is not the same as authorship.
AI-Generated Van Gogh Art Rejected by US Copyright Office
Artist Ankit Sahni’s copyright application for an AI-generated artwork that combined the artist’s original base photograph with Vincent Van Gogh’s Starry Night painting has been rejected again by the US Copyright Office. According to the US office, the new artwork “lacks the human authorship necessary to support a copyright claim” and “human authorship cannot be distinguished or separated from the final work produced by the computer program.” The artist originally tried to register the artwork in 2021 and described the image as a photograph, 2-D artwork, and the RAGHAV AI Painting App as the author of 2-D artwork.
Ginger Liu is the founder of Ginger Media & Entertainment, a Ph.D. Researcher in artificial intelligence and visual arts media — specifically death tech, digital afterlife, AI death and grief practices, AI photography, entertainment, security, and policy, and an author, writer, artist photographer, and filmmaker. Listen to the Podcast — The Digital Afterlife of Grief.
