Media and entertainment companies are now exploring fine-tuning video generation models to create custom model versions for their own internal use, including potentially on specific productions. Fine-tuning refers to a process of training a pre-trained AI model on a curated dataset to create a smaller new model, which is then capable of producing more specific kinds of outputs.
Source: Fine-Tuning AI Video Models Getting Early Interest From Film & TV Studios
Less than a year ago, the drive for AI regulation was gaining significant momentum, marked by key milestones such as the AI Safety Summit in the U.K., the Biden Administration’s AI Executive Order, and the EU AI Act. However, a recent judicial decision and potential political shifts are leading to more uncertainty about the future of AI regulation in the U.S.
Today, Meta said it is releasing Llama 3.1 405B, a model containing 405 billion parameters. Parameters roughly correspond to a model’s problem-solving skills, and models with more parameters generally perform better than those with fewer parameters. At 405 billion parameters, Llama 3.1 405B isn’t the absolute largest open source model out there, but it’s the biggest in recent years.
DJ Calvin Harris sold his catalog that includes more than 150 songs and collaborations with some of the world’s biggest stars, including Rihanna, Dua Lipa, Sam Smith, Frank Ocean, Ariana Grande, Big Sean, Khalid, Pharrell Williams, and Travis Scott. Sony/ATV continued administration services for the catalog. Now more than four years later, Shamrock Capital Advisors has engineered a large deal including film, television show, and song rights from Vine Alternative Investments in New York
Meta says it won’t be launching its upcoming multimodal AI model — capable of handling video, audio, images, and text — in the European Union, citing regulatory concerns. The decision will prevent European companies from using the multimodal model, despite it being released under an open license. “We will release a multimodal Llama model over the coming months, but not in the EU due to the unpredictable nature of the European regulatory environment,” Meta spokesperson Kate McLaughlin said.
Source: Meta won’t release its multimodal Llama AI model in the EU
Artificial intelligence-powered search engines provided by U.S. tech giants like Google LLC and Microsoft Corp. likely infringe on copyright, an association run by Japanese mass media says. The Japan Newspaper Publishers and Editors Association, in a statement, called for companies operating such services to obtain consent from news organizations as search responses often resemble articles that are sourced without permission.
Source: Japanese media say AI search infringes copyright, urge legal reform
Under the proposed law, this “content provenance information” would be embedded in digital forms of copyrighted material, and it would be unlawful to remove it or tamper with it, except in very limited cases where platforms are carrying out research to improve security. It would also be unlawful for anyone to use any material with “content provenance information” to train AI, or to create AI-generated content, “unless such person obtains the express, informed consent of the person who owns the covered content.”
The threat of deepfakes is growing as the AI tools to create them become widely accessible. There was a 245% increase in deepfakes worldwide from 2023 to 2024, an uptick spurred in part by coming election cycles, according to verification provider Sumsub. The trend, unsurprisingly, has been a windfall for companies marketing tools to defend against deepfakes and technologies used to produce them.
Source: Deepfake-detecting firm Pindrop lands $100M loan to grow its offerings
AI companies are generally secretive about their sources of training data, but an investigation by Proof News found some of the wealthiest AI companies in the world have used material from thousands of YouTube videos to train AI. Companies did so despite YouTube’s rules against harvesting materials from the platform without permission.
Source: Apple, Nvidia, Anthropic Used Thousands of Swiped YouTube Videos to Train AI