Anthropic wins fair use case as OpenAI faces copyright lawsuits

Key Takeaways

• OpenAI faces lawsuits from Encyclopedia Britannica and Merriam-Webster, alleging copyright infringement for using nearly 100,000 articles to train models like ChatGPT without permission.
• U.S. courts in 2025 began shaping AI copyright law, with a ruling in Bartz et al. v. Anthropic finding fair use for lawfully acquired books, while Andersen v. Stability AI and Kadrey et al. v. Meta address other data sourcing issues.
• Private equity firms are partnering with major AI companies, including OpenAI and Anthropic, to form enterprise AI consulting arms, serving as investors and initial customers to integrate AI into portfolio companies.
• Google is expanding its 'Personal Intelligence' features across the U.S. in Search, Gemini app, and Chrome, connecting user-controlled Google apps like Gmail and Photos for personalized responses without direct data training.
• AI is being adopted in education, with Telangana, India, planning AI tutors for specialized courses, and Bucks County, Pennsylvania, schools piloting programs like Central Bucks' use of Microsoft Copilot.
• Distance Education for Africa trained 1,068 participants from 45 African countries in practical AI literacy, emphasizing AI as a collaborator in the changing job market.
• The European Writers' Council urged WIPO to enforce the ART principle (Authorization, Remuneration, Transparency) and mandatory labeling for AI-generated content, citing an existential threat to authors.
• AI shopping assistants like Google's Gemini and OpenAI's ChatGPT demonstrate limitations, such as text-only recommendations or inability to complete direct purchases, despite offering personalized experiences.
• Former Google CEO Eric Schmidt predicts top programmers will become more valuable in the AI era due to their unique mathematical reasoning skills, leading to a more specialized job market.
• Netflix co-CEO Ted Sarandos advocates for using AI to enhance content quality rather than merely reducing costs, viewing it as a tool for creators while maintaining the essential role of human talent.

AI training data lawsuits shape copyright law in 2025

In 2025, U.S. courts made key decisions on whether using copyrighted material to train AI is legal. The cases involved companies like Anthropic and Meta, and addressed issues like fair use and market harm. A ruling in Bartz et al. v. Anthropic found that training AI on lawfully acquired books was fair use. However, cases like Kadrey et al. v. Meta highlighted risks with improperly sourced data. The Andersen v. Stability AI case is ongoing, focusing on image generation models trained on scraped artworks. These rulings provide early guidance for AI developers and content creators.

Britannica sues OpenAI over AI training data

Encyclopedia Britannica and Merriam-Webster are suing OpenAI, alleging that ChatGPT and other AI models were trained using their copyrighted texts without permission. The lawsuit claims nearly 100,000 articles were used and that ChatGPT can reproduce content verbatim, potentially harming Britannica's website traffic. Britannica also accuses OpenAI of trademark infringement. OpenAI argues its training data is publicly available and falls under fair use. The publishers seek financial compensation and an end to the alleged use of their material.

Britannica sues OpenAI over ChatGPT training data

Encyclopedia Britannica has filed a lawsuit against OpenAI, claiming that tens of thousands of its copyrighted articles were used without permission to train AI models like ChatGPT. The lawsuit argues this constitutes copyright infringement and could harm Britannica's business by allowing AI to reproduce content. This case is part of a larger trend of lawsuits against AI companies over training data. Publishers, authors, and media organizations are increasingly challenging the use of their work without consent, questioning whether it constitutes fair use or infringement.

Britannica and Merriam-Webster sue OpenAI for AI training

Encyclopedia Britannica, along with its subsidiary Merriam-Webster, has sued OpenAI for copyright infringement. They claim nearly 100,000 copyrighted articles were used without authorization to train OpenAI's large language models. The publishers also allege that OpenAI's systems reproduce their content and raise trademark infringement concerns. This lawsuit joins others from media organizations like The New York Times against OpenAI, as courts examine how copyright law applies to AI training and outputs.

AI shopping assistants like ChatGPT and Gemini offer mixed results

The Globe and Mail tested AI shopping assistants ChatGPT and Gemini to see if they could help users buy sweaters. Reporters found that while AI promises a personalized shopping experience, current tools have limitations. Gemini provided text-only recommendations, requiring users to search separately for product images. ChatGPT sometimes required very specific prompts and couldn't complete purchases directly in Canada. Both platforms showed potential but also highlighted areas needing improvement for seamless online shopping.

Google expands Personal Intelligence features in the US

Google is expanding its 'Personal Intelligence' features across the U.S. in its Search, Gemini app, and Chrome. This feature connects Google apps like Gmail and Photos to provide personalized responses, such as tailored shopping recommendations or travel itineraries. Users have control over which apps connect and can turn these features on or off. Google emphasizes that this is designed with privacy in mind, and the system does not train directly on user data from Gmail or Photos. The service is available for free to personal Google account users.

Telangana uses AI tutors for specialized courses

Telangana, India, is planning to use artificial intelligence (AI) tutors to teach new, specialized courses. This initiative aims to address a shortage of trained faculty in fields like maritime technology and air traffic management. The state government sees AI as a way to provide quality education and bridge the gap in specialized training. AI tutors are expected to offer consistent and accessible learning experiences, complementing existing educational resources.

1,068 Africans trained in AI skills for the future

Distance Education for Africa (DE Africa) recently completed a seven-week program called 'Your Journey in the AI Economy' for 1,068 participants from 45 African countries. The course focused on practical AI literacy to help people adapt to the changing job market. Learners from diverse backgrounds, including Botswana, Kenya, and Uganda, participated. The program emphasized judgment and decision-making, treating AI as a collaborator rather than just a tool. Despite challenges like internet issues, participants found the course valuable for professional development and applying AI in their work.

Authors urge WIPO to respect AI ART principle

The European Writers' Council (EWC) intervened at the World Intellectual Property Organization's (WIPO) Artificial Intelligence Infrastructure Interchange (AIII) launch. EWC highlighted that book authors are significantly impacted by AI development and products. They urged WIPO to enforce the ART principle: Authorization, Remuneration, and Transparency. EWC called for mandatory, multi-layered labeling of AI-generated content to ensure provenance and proper licensing. They stressed that AI poses an existential threat to authors' creativity and work.

Cursor vs. GitHub Copilot: Which AI coding assistant is best?

This article compares Cursor, an AI-first Integrated Development Environment (IDE), with GitHub Copilot, an AI coding assistant plugin. Cursor aims to reinvent the IDE with deep AI integration, while Copilot enhances existing IDEs. Cursor offers stronger codebase context and more native AI features, whereas Copilot boasts broader compatibility and a mature autocompletion system. The choice depends on whether developers prefer a new AI-centric environment like Cursor or an integrated assistant like Copilot within their current tools.

Top programmers will be more valuable in AI era, says ex-Google CEO

Former Google CEO Eric Schmidt believes top programmers will become even more valuable as AI advances. He argues that their unique mathematical reasoning skills are crucial for controlling AI systems. Schmidt predicts a market shift where a few large companies and many small ones dominate, with fewer programmers needed overall. He suggests the best programmers will leverage AI's parallelization capabilities, leading to a bifurcation in the programming job market with elite talent becoming rarer and more sought after.

Private equity firms partner with AI giants

Private equity firms are forming partnerships with major AI companies like OpenAI and Anthropic. These collaborations aim to create enterprise AI consulting arms, where PE firms act as minority investors and initial customers. This move allows PE firms to help their portfolio companies integrate AI and address the threat AI poses to their investments. For AI companies, these partnerships offer a faster way to expand into the enterprise market and test new business models.

Bucks County schools pilot AI, create policies

Nine school districts in Bucks County, Pennsylvania, are exploring artificial intelligence (AI) through pilot programs, training, and policy development. Approaches vary, with some districts forming 'AI think tanks' to create guidance, while others are launching specific AI pilot programs for students, like Central Bucks' use of Microsoft Copilot. Pennsbury School District has already implemented a formal AI policy with assignment scales. The University of Pennsylvania's Graduate School of Education is supporting these efforts through its Pioneering AI in School Systems program.

Netflix uses AI to make content better not cheaper

Netflix co-CEO Ted Sarandos believes AI should be used to make content 'better,' not just faster or cheaper, especially in the competitive media landscape. He sees AI as a tool for creators, similar to evolving production technologies. While AI can assist with editing and production, Sarandos emphasizes that human talent like writers and actors remains essential. He notes AI struggles with nuanced performance, particularly in voice acting, but sees potential for AI to help recreate lines post-production to improve films.

Sources

• Copyright Law in 2025: Courts begin to draw lines around AI training, piracy, and market harm
• Encyclopedia Britannica sues OpenAI over AI training
• Encyclopedia Britannica just sued OpenAI over ChatGPT — here’s why AI training is under fire (again)
• Britannica and Merriam-Webster File Lawsuit Against OpenAI Over AI Training and Content Use
• We tried to buy sweaters with AI. Here’s how it went
• Bringing the power of Personal Intelligence to more people
• Telangana to rely on AI tutors to teach new courses
• DeAfrica’s AI Training: 1,068 Learners In 45 African Countries.
• WIPO's AIII Launch: EWC's intervention and appeal to always respect the ART principle - EWC
• Cursor vs. GitHub Copilot (2026): Qual É O Melhor Assistente De Programação IA?
• Top Programmers Will Be More Valuable In The AI Era: Former Google CEO Eric Schmidt
• Private equity is partnering with Big AI
• Bucks County Schools Roll Out AI Pilots, Training and Policies for the Classroom
• How Netflix thinks AI can help it fight off rivals