📜 OpenAI Wins Copyright Case

OpenAI's victory in an initial copyright case could set a friendly precedent for AI training and fair use, dealing a blow to publisher efforts that fought OpenAI through the last year.

Author

Jack Lajoie
November 08, 2024 • read time ~ 8 minutes

The Daily Current ⚡️

Welcome to the creatives, builders, pioneers, and thought leaders ever driving further into the liminal space.

OpenAI's recent legal victory in a copyright case marks a significant step forward for AI development, potentially establishing a new precedent for fair use in AI training. Alongside this, ByteDance's advancements in AI portrait animation and Baidu's upcoming AI smart glasses underscore the rapid integration of AI in creative and consumer technologies. Meanwhile, Perplexity AI's success during the U.S. election highlights its potential as a leader in AI-driven information delivery, and tech CEOs' strategic engagement with Donald Trump signals a shift in the industry's political maneuvering

 🔌 Plug Into These Headlines:

A federal court has ruled in favor of OpenAI in a copyright dispute with news outlets, determining that the company’s use of news articles for AI training is permissible under fair use doctrine. The decision, made by Judge Victor Marrero in Manhattan, supports the argument that AI training is transformative and does not compete with or diminish the value of the original works. This ruling may facilitate continued AI development using various copyrighted sources for training data.

  • Judge Victor Marrero dismissed the lawsuit filed by news organizations against OpenAI and Microsoft on November 7, 2024.

  • The ruling states that using news articles to train AI models falls under the “fair use” doctrine of copyright law.

  • The decision supports OpenAI’s argument that their use of news content is transformative and doesn’t compete with original works.

🔁 While still too early to call, this ruling couple with Trump’s recent election may truly force news organizations to reconsider their business models and explore new ways to monetize their content in an AI-driven world.

X-Portrait 2 enhances portrait animation capabilities by generating realistic character animations using only a static portrait and a driving video. The technology employs a sophisticated expression encoder trained on large-scale datasets, combined with generative diffusion models. This approach enables the transfer of nuanced facial expressions and challenging emotions, outperforming other methods in capturing fast head movements and minuscule expression changes.

  • The technology combines the encoder with powerful generative diffusion models for fluid and expressive video generation.

  • X-Portrait 2 can transfer subtle facial expressions and challenging expressions like pouting and tongue-out.

  • The model achieves strong disentanglement of appearance and motion during training.

💻 The development of X-Portrait 2 underscores the growing importance of machine learning in content creation, hinting at a shift in the skills required for future animation professionals, especially as it comes from the TikTok parent company.

At the annual Baidu World event in Shanghai, China’s leading search engine company will unveil its latest AI hardware offering: smart glasses with an integrated AI assistant. This product aims to compete with Meta’s successful Ray-Ban smart glasses, featuring similar capabilities such as built-in cameras and voice interactions.

  • Baidu will unveil the AI-powered smart glasses at Baidu World event in Shanghai next week

  • The glasses will feature built-in cameras for capturing photos and videos

  • Voice interactions will be supported using Baidu’s Ernie foundation model

  • This product is seen as a direct rival to Meta’s Ray-Ban smart glasses

  • The announcement demonstrates Baidu’s expansion into AI-powered hardware

👓 The introduction of Baidu’s AI smart glasses further suggests that voice-activated, camera-equipped eyewear could become a mainstream consumer product category in the near future.

In a high-stakes test of AI capabilities during the 2024 U.S. Presidential election, Perplexity AI successfully delivered timely and mostly accurate election-related information. Unlike competitors such as OpenAI’s ChatGPT and Google’s Gemini, Perplexity took a calculated risk by launching an election information hub, which paid off as users turned to the platform for real-time updates and insights.

  • Perplexity AI successfully provided real-time election insights and maps, while competitors (those that tried — like xAI’s Grok) failed or refused to answer election-related questions.

  • Perplexity’s approach combined traditional visual election charts with AI-generated answers, resulting in a comprehensive and reliable information source for users.

  • While some minor issues and hallucinations occurred, Perplexity’s overall performance was praised for its accuracy and timeliness during the election night.

👏 The success of Perplexity during this high-stakes event positions it as a leader in the AI-powered information delivery space for future elections.

Tech industry leaders have shifted their approach to Donald Trump ahead of the 2024 election, proactively engaging with him instead of keeping their distance as they did in 2016. Executives like Mark Zuckerberg, Tim Cook, and Sundar Pichai have made direct contact with Trump, aiming to establish personal relationships that could influence future policies affecting their companies. This strategy reflects lessons learned from Trump’s first term and recognizes his transactional approach to politics.

  • With speculation about potential changes in leadership at regulatory agencies like the FTC under a new Trump administration, it’s likely tech CEOs wanted to be proactive in forging inroads to weigh in.

  • This NYT article suggests that tech executives see value in presenting a united front when dealing with Trump.

  • The strategy reflects a recognition of Trump’s tendency to be influenced by personal relationships and direct communication.

🤔 Leave it to the New York Times to wait until after the election to publish something like this, which they must have known about during the summer at least. It’s not surprising to anyone with an eye to the potential that lower regulations under Trump could hold for the tech industry at this juncture.