June 28, 2025

news   Read News
Meta, Anthropic AI Legal Wins; Salesforce Job Cuts, Scale AI $14B

Recent developments in the AI space cover a wide range of topics, from copyright battles to job displacement and regulatory concerns. Meta and Anthropic have seen some legal wins regarding the use of copyrighted material for AI training, though the broader debate about compensating copyright holders continues, potentially heading to the Supreme Court. Specifically, a judge dismissed a lawsuit against Meta, stating the authors made incorrect arguments, while Anthropic won a copyright lawsuit but still faces issues related to illegally obtaining books for training its Claude chatbot. Meanwhile, AI's increasing capabilities are impacting the job market. Salesforce CEO Marc Benioff noted that AI now handles up to 50% of the company's workload, leading to around 1,000 job cuts, though Salesforce is also hiring for AI-related roles. Mark Zuckerberg is making significant investments in AI, including a $14 billion bet on Scale AI, which helps improve AI models. Google stopped using Scale AI after Meta's investment was announced. On the energy front, Texas Tech and Fermi America are collaborating on a massive AI energy campus powered by a mix of nuclear, solar, wind, and natural gas, aiming for the first gigawatt of capacity by the end of 2026. In AI applications, a new AI Assistant (Beta) is designed to simplify data center management. However, AI models are still struggling with certain aspects of human language, as evidenced by their difficulty in understanding Gen Alpha slang. Creative Commons is developing CC signals to allow creators to specify whether they want their work used for AI training and under what conditions. Finally, the Trump administration's proposed 'Big Beautiful Bill' could potentially centralize AI regulation at the federal level, impacting states like North Carolina, and a new AI platform called ICERAID.us is gamifying US border monitoring, raising ethical concerns.

Key Takeaways

  • Meta and Anthropic won copyright lawsuits regarding AI training, but the legal battle over compensating copyright holders is ongoing.
  • Anthropic faces a trial for illegally obtaining over 7 million books to train its Claude chatbot, despite winning a copyright lawsuit.
  • Salesforce CEO Marc Benioff reports AI handles 50% of the company's work, resulting in 1,000 job cuts but also new AI-related hires.
  • Mark Zuckerberg is investing $14 billion in Scale AI, an AI startup that improves AI models through human feedback.
  • Google stopped using Scale AI after Meta's investment in the company was announced.
  • Texas Tech and Fermi America are building a 5,800-acre AI energy campus powered by nuclear, solar, wind, and natural gas, with the first gigawatt expected by 2026.
  • A new AI Assistant (Beta) aims to streamline data center management by helping users find features and run reports.
  • AI models like GPT-4, Claude, Gemini, and Llama 3 struggle with understanding Gen Alpha slang, posing challenges for online safety.
  • Creative Commons is developing CC signals to allow creators to control the use of their work in AI training, including specifying conditions like credit or payment.
  • The Trump administration's proposed 'Big Beautiful Bill' could centralize AI regulation at the federal level, potentially preventing states like North Carolina from enacting their own AI laws.

Meta and Anthropic win AI copyright battles but war is not over

Meta and Anthropic won legal cases about using copyrighted books to train their AI. However, the bigger fight about paying copyright holders is still ongoing. A judge said Anthropic can use books they paid for, but must face claims about using pirated books. Another judge said authors suing Meta made incorrect arguments. Experts think these issues will eventually go to the Supreme Court.

Anthropic wins copyright lawsuit but faces piracy issues

Anthropic, an AI company, won a lawsuit about copyright infringement for training its Claude chatbot. A judge said using copyrighted material to train AI is considered fair use. However, Anthropic also downloaded over 7 million books from online shadow libraries without permission. The judge said this was copyright infringement and could cost Anthropic millions of dollars in damages. Another case against Meta had a similar ruling but gave plaintiffs a roadmap to win future claims.

Judge dismisses authors' lawsuit against Meta over AI training

A judge dismissed a lawsuit from authors against Meta, the parent company of Facebook. The authors claimed Meta used their copyrighted books to train its AI technology. The judge stated the authors made the wrong arguments, but this doesn't mean Meta's use of copyrighted material is lawful. Another judge ruled that Anthropic, an AI company, didn't break the law by training its chatbot on copyrighted books, but must still go to trial for illegally obtaining the books.

AI wins early battle with Hollywood over copyright use

AI companies are gaining ground in their fight with Hollywood over copyrighted material. A judge ruled that Anthropic can train its AI models on copyrighted work if obtained legally. This decision may make AI companies less likely to make deals with Hollywood studios. Experts say AI companies can now buy content like scripts and use them for training as long as they keep their receipts. Disney is suing Midjourney for allowing users to create copies of characters like Homer Simpson.

Lost your job to AI? What to do next

AI is becoming more common in Silicon Valley, causing concern about job losses. Salesforce CEO Marc Benioff says AI has helped cut about 1,000 jobs. However, Salesforce has also hired more people to sell and maintain AI products. One tech executive, Mark Quinn, lost his job to AI but found a new role in the same industry. He now works for Pearl, an AI search platform.

Salesforce CEO says AI handles half the company's work

Salesforce CEO Marc Benioff says AI now handles up to 50% of the company's workload. This announcement has made some job seekers nervous. Benioff admits AI has led to about 1,000 job cuts. However, the company has also hired more people to sell and maintain its AI products.

Gen Alpha slang confuses parents and AI models

Gen Alpha slang is hard for both parents and AI to understand. A study showed that AI models like GPT-4, Claude, Gemini, and Llama 3 struggled with Gen Alpha phrases. AI had trouble understanding masked harassment, which is a concern for online safety. Parents scored about the same as the best AI model in understanding the slang. Gen Alpha's online language changes quickly, making it hard for adults to keep up.

Zuckerberg bets $14 billion on AI startup Scale AI

Mark Zuckerberg is investing heavily in Scale AI, an AI startup. Scale AI helps improve AI models by having humans rewrite chatbot responses. This expertise is valuable to Zuckerberg as he wants to release new AI models. Google stopped using Scale AI after Meta's investment was announced. Scale AI uses a platform called Remotasks with over 240,000 contributors to train AI models.

Texas Tech and Fermi America plan huge AI energy campus

Texas Tech University System and Fermi America are partnering to build a large AI energy campus near Amarillo. The campus will use nuclear, solar, wind, and natural gas to power AI data centers. The project aims to meet the growing energy needs for AI and boost U.S. technology competitiveness. The campus will cover 5,800 acres and use up to 11 gigawatts of electricity. The first gigawatt of capacity is expected by the end of 2026.

New AI Assistant makes data center management easier

A new AI Assistant (Beta) is available to help users find features and run reports. The AI Assistant can explain how things work and help users get things done faster. The search system has been improved to be faster and easier to use. Users now have more control over asset placement with horizontal and vertical offsets. Other enhancements include better insight into asset lifecycle states and more precise documentation of network connections.

BizLink expands for AI hardware growth with new sites

BizLink Group, a global provider of interconnect solutions, opened a new facility in Tainan. This shows the company's focus on local research and development and production. Chairman Roger Liang said the company is sticking to its goal of double-digit revenue growth. The article requires a paid subscription to read more.

AI turns US border enforcement into a game

A new AI platform called ICERAID.us is encouraging people to monitor the US border. It offers cryptocurrency rewards for uploading photos of suspicious activity. The platform is led by Enrique Tarrio, who describes himself as an ICE Raid Czar. Critics say this gamifies surveillance and could lead to violence. The platform's URL is similar to a website that tracks ICE raids, which may confuse people.

Federal government may soon regulate AI, impacting North Carolina

The Trump administration's proposed 'Big Beautiful Bill' could stop states like North Carolina from regulating AI for 10 years. This would leave AI regulation to the federal government. North Carolina is currently debating bills like House Bill 375, which would require disclaimers on AI-generated political ads. Supporters of the bill hope it will create uniform federal AI regulation. However, some worry Congress will struggle to agree on nationwide AI regulations.

Creators may soon get paid for AI training use

Creative Commons is developing a new system called CC signals that lets creators say if they want their work used for AI training. Creators can say yes or no and add conditions like getting credit or payment. AI developers can scan content for CC signals to see the creator's preferences. The CC signals are not enforceable yet but act as social and ethical markers. Some platforms are already making changes to control AI data scraping of their users' content.

Sources

AI Copyright Lawsuit Meta Anthropic Fair Use Piracy AI Training Job Losses Salesforce Gen Alpha Slang AI Models Scale AI Mark Zuckerberg AI Energy Campus Data Center Management AI Assistant BizLink Group US Border Enforcement ICERAID.us AI Regulation North Carolina Creative Commons CC Signals AI Data Scraping GPT-4 Claude Gemini Llama 3