openai, microsoft and meta Updates

The artificial intelligence sector is currently navigating significant legal challenges, evolving market dynamics, and expanding applications across various industries. The New York Times recently filed a lawsuit against AI startup Perplexity AI in federal court, accusing it of copyright infringement by using its content without permission and creating "hallucinations" wrongly linked to the newspaper's trademarks. This action follows previous legal challenges by The Times against OpenAI and Microsoft, and similar lawsuits against Perplexity by other publishers like Dow Jones, the New York Post, and the Chicago Tribune. Perplexity, which claims it only indexes web pages and provides citations, has dismissed these lawsuits as outdated tactics, even as PayPal integrates its payment system into AI-native checkout experiences with Perplexity and others. In contrast to the legal disputes over content use, Meta has adopted a different strategy, securing multi-year content licensing deals with major publishers including CNN, Fox News, People Inc., and USA Today Co. These agreements allow Meta to use both new and archival content to train its large language model, Llama. Meanwhile, Salesforce CEO Marc Benioff suggests that large language models from companies like OpenAI, Google, and Anthropic are becoming commoditized, much like disk drives, with businesses prioritizing cost efficiency and flexibility. Google's Gemini 3 Pro, however, demonstrates advanced capabilities as a multimodal model, setting new records in vision AI for complex visual and spatial reasoning tasks. The infrastructure supporting AI development is also seeing shifts, as a Dell Oro Group report from December 5, 2025, indicates Ethernet now accounts for over two-thirds of data center switch sales in AI back-end networks, a significant increase from less than half the previous year. This growth outpaces InfiniBand, despite NVIDIA's Blackwell Ultra platform boosting InfiniBand sales, driven by strong demand from major AI cluster builders such as Google, Amazon, Microsoft, and Meta. Accton, Arista, and Cisco have gained substantial market share in this segment, with most shipments being 800 Gbps switches. AMD CEO Lisa Su maintains that the current investment in AI does not constitute an "AI bubble," emphasizing that a bubble occurs when asset prices far exceed actual value due to speculation. Beyond corporate developments, AI is making inroads into education and sparking philosophical debates. The TALIS 2024 report reveals that 41% of teachers in OECD countries now use AI in their teaching, primarily for learning topics and lesson planning, though concerns about cheating and incorrect suggestions persist. In response to the growing importance of AI, the UW Board of Regents approved the creation of a new College of Computing and Artificial Intelligence (CAI) at UW-Madison, set to begin operations in July 2026. This marks the university's first new college in over 50 years. However, the role of AI in creative fields remains contentious, with some arguing that true art requires human experience, effort, and passion, qualities that artificial intelligence does not possess.

Key Takeaways

• The New York Times is suing AI startup Perplexity AI for copyright infringement, illegal content copying, and trademark violation, alleging it uses Times content without permission and generates "hallucinations" linked to the newspaper.
• Other publishers, including Dow Jones, New York Post, and Chicago Tribune, have also filed similar lawsuits against Perplexity AI.
• Meta has opted for content licensing, signing multi-year deals with major publishers like CNN, Fox News, and USA Today Co. to train its Llama large language model.
• Salesforce CEO Marc Benioff believes large language models from companies like OpenAI, Google, and Anthropic are becoming commoditized infrastructure, with businesses seeking the cheapest and best options.
• Google's Gemini 3 Pro is a highly capable multimodal model, setting new records in vision AI for understanding documents, spatial relationships, screens, and videos.
• Ethernet now comprises over two-thirds of data center switch sales in AI back-end networks, surpassing InfiniBand, driven by demand from Google, Amazon, Microsoft, and Meta for 800 Gbps switches.
• AMD CEO Lisa Su states there is no "AI bubble," defining a bubble as asset prices exceeding actual value due to speculation.
• 41% of teachers across OECD countries use AI in their teaching, mainly for learning and lesson planning, according to the TALIS 2024 report, though concerns about cheating and incorrect suggestions are common.
• UW-Madison received approval to establish a new College of Computing and Artificial Intelligence (CAI), expected to begin operating in July 2026, marking its first new college in over 50 years.
• PayPal is integrating its payment system into AI-native checkout experiences with partners like Perplexity, aiming to facilitate

  New York Times Sues Perplexity AI for Copyright Violation

  The New York Times filed a lawsuit against AI startup Perplexity on Friday in federal court in New York. The Times claims Perplexity repeatedly violated its copyrights by using its content without permission. Perplexity's AI search engine allegedly copied large parts, sometimes entire articles, directly competing with the newspaper. The Times had contacted Perplexity many times over 18 months before filing this suit. This is the latest in several lawsuits against AI companies, including a similar one by Dow Jones against Perplexity and a previous one by The Times against OpenAI and Microsoft.

  New York Times Sues Perplexity for Illegal Content Copying

  The New York Times sued AI startup Perplexity AI for illegally copying millions of its articles. The lawsuit also claims Perplexity violated the Times's trademarks by creating false content, called "hallucinations," and wrongly linking them to the newspaper. Perplexity's business model relies on taking content, including paywalled material, to power its AI products. Other publishers like Dow Jones and New York Post have also sued Perplexity for similar reasons. Cloudflare and Reddit have also made accusations against the San Francisco-based company.

  New York Times Sues Perplexity AI Over Content Use

  The New York Times sued Perplexity AI on December 5, accusing the startup of illegally copying its content. The Times claims Perplexity's AI products create false information, or "hallucinations," and wrongly link them to the newspaper's trademarks. Perplexity's business model reportedly scrapes and copies content, including paywalled material. NYT spokesperson Graham James stated they object to Perplexity's unlicensed use of their content. Perplexity's Jesse Dwyer called the lawsuits an outdated tactic, while the company previously claimed it only indexes web pages and provides citations. The Times seeks damages and a court order to stop the alleged unauthorized use.

  New York Times Sues Perplexity AI for Illegal Article Use

  The New York Times has sued AI startup Perplexity AI in Manhattan federal court. The lawsuit claims Perplexity illegally copied thousands of the Times's articles without permission to train its AI chatbot. The Times states that Perplexity's chatbot showed its content in user replies, acting like a search engine that avoids the original publisher. This legal action is part of a growing trend of media companies challenging AI firms over copyright issues. Perplexity has denied similar claims before, saying such lawsuits are old-fashioned and its technology does not break copyright laws.

  New York Times Sues Perplexity for AI Copyright Breach

  The New York Times filed a lawsuit against AI startup Perplexity on Friday for copyright infringement. The Times stated it had contacted Perplexity many times about the issue. This lawsuit follows similar actions, including one by the Chicago Tribune on Thursday and previous suits by Dow Jones, owner of the Wall Street Journal and New York Post. Perplexity has a Publisher's Program that shares ad revenue with news outlets like Gannett and Time. Graham James, a spokesperson for The Times, said they object to Perplexity's unlicensed use of their content.

  UW Madison Approved for New AI Computing College

  The UW Board of Regents approved UW-Madison's plan to create a new College of Computing and Artificial Intelligence, or CAI. School officials announced this on Thursday and will begin planning for the college. An official announcement about the new college will come in the spring. The CAI is expected to start operating in July 2026. Chancellor Jennifer Mnookin said this new college will prepare students for a future where computing and technology meet.

  UW Regents Approve New AI Computing College

  The UW Board of Regents approved creating a new College of Computing and Artificial Intelligence, or CAI, on December 4, 2025. This involves reorganizing the School of Computer, Data and Information Sciences, which will separate from the College of Letters and Science in summer 2026. The new college will combine computer sciences, the Information School, and statistics, focusing on a human-centered approach. It will serve as a central point for AI education across campus. This marks the first new college at UW in over 50 years.

  Teachers Use AI More in Class Says TALIS 2024 Report

  The TALIS 2024 report shows that 41% of teachers across OECD countries now use artificial intelligence in their teaching. Most teachers use AI to learn about topics and create lesson plans, but fewer use it for grading or checking student performance. While 52% believe AI can help tailor learning, 72% worry it lets students cheat and 66% fear incorrect suggestions. Many teachers who do not use AI lack the necessary skills, but only 48% think AI should not be used in teaching. Education systems are responding, with 38% of teachers receiving AI training in the past year, showing a positive link between training and AI use. AI also has the potential to reduce teachers' administrative work, which 52% find very stressful.

  True Art Needs Human Experience Not Artificial Intelligence

  This column argues that artificial intelligence cannot create true art because real art comes from human experience. The author believes it is important to keep AI separate from art creation and writing. Hayao Miyazaki of Studio Ghibli once called an AI-generated creature "an insult to life itself," a sentiment the author feels is even truer today. While some believe AI art lets everyone be an artist, the author states that art is about effort, passion, and humanity, none of which AI possesses. Art is the purest form of human expression, helping us connect and understand each other. The author warns against letting AI remove humanity from the arts, fearing a future of bland and uninspired creations.

  Salesforce CEO Says AI Models Are Now Like Disk Drives

  Salesforce CEO Marc Benioff believes that large language models, or LLMs, have become a commodity, much like disk drives. He stated on X that LLMs are now interchangeable infrastructure, where companies will choose the cheapest and best option. Benioff himself recently switched from ChatGPT to Gemini 3, showing how easy it is to change models. This trend suggests that major AI models from companies like OpenAI, Google, and Anthropic are becoming very similar in their abilities. While some argue that future breakthroughs could lead to more unique models, Benioff's view highlights the current market where businesses seek flexibility and cost efficiency.

  PayPal Integrates AI Checkout with Perplexity for New Shopping

  PayPal is changing its investment story by integrating its payment system into AI-native checkout experiences, like those with Perplexity, Ashley, and Newegg. This move allows people to complete purchases directly from AI search results, moving from asking questions to paying within conversational tools. PayPal aims to be a key part of "agentic commerce," where AI agents help consumers find, compare, and buy products. This strategy could boost PayPal's growth and strengthen its role as a vital payment provider in the future of online shopping. However, its success depends on how widely AI shopping tools are used and if customers are willing to buy through these new interfaces.

  Ethernet Dominates AI Network Switch Sales in 3Q 2025

  A new report from Dell Oro Group on December 5, 2025, shows that Ethernet now makes up over two-thirds of data center switch sales in AI back-end networks. This is a big jump from less than half last year, indicating a clear shift from InfiniBand. While NVIDIA's Blackwell Ultra platform boosted InfiniBand sales, Ethernet's growth is faster due to high demand from major AI cluster builders like Google, Amazon, Microsoft, and Meta. Accton, Arista, and Cisco gained the most market share, while Celestica and NVIDIA together hold nearly 50% of the Ethernet segment. Most of these Ethernet switch shipments were 800 Gbps switches.

  AMD CEO Lisa Su Says No AI Bubble Exists

  AMD CEO Lisa Su firmly stated in a PC Gamer interview that she does not believe there is an "AI bubble." She explained that a bubble occurs when asset prices far exceed actual value due to speculation. While many companies are investing heavily in AI, opinions are split on whether it's a bubble. Su's company, AMD, has seen its stock price rise significantly, similar to Nvidia, which has also benefited greatly from the AI boom. She suggests that whether AI is a bubble depends on if AI companies can truly deliver on their promises and achieve widespread, long-term adoption.

  Meta Strikes AI Content Licensing Deals with Major Publishers

  Meta has joined the AI content licensing market by signing multi-year deals with several major publishers, including CNN, Fox News, People Inc., and USA Today Co. These agreements allow Meta to use new and old content from these sources to train its large language model, Llama. While the financial terms are not public, this move shows Meta is formalizing how it gets news content for AI development. Publishers see these deals as a win, confirming the value of their content and the possibility of fair licensing in the competitive AI landscape.

  Gemini 3 Pro Leads in Vision AI Capabilities

  Google's Gemini 3 Pro is a highly capable multimodal model that performs well in understanding documents, spatial relationships, screens, and videos. It represents a big step forward from simple recognition to true visual and spatial reasoning. The model sets new records on vision tests like MMMU Pro and Video MMMU for complex visual tasks. Gemini 3 Pro can accurately detect and recognize text, tables, and math formulas in messy documents. It also has strong spatial understanding, allowing it to identify objects and point to specific locations in images with great accuracy.

  Sources

  • New York Times Sues A.I. Start-Up Perplexity Over Use of Copyrighted Work
  • New York Times sues AI startup for ‘illegal’ copying of millions of articles
  • New York Times sues Perplexity AI for 'illegal' copying of content
  • New York Times sues Perplexity AI for 'illegal' copying of content
  • New York Times sues Perplexity for copyright infringement with AI
  • New College of Computing and Artificial Intelligence approved for UW–Madison
  • UW Board of Regents approves new College of Computing and Artificial Intelligence
  • Using Artificial Intelligence: Takeaways from TALIS 2024
  • COLUMN | Artificial intelligence cannot create true art
  • LLMs Are Now A Commodity, Like Disk Drives: Salesforce CEO Marc Benioff
  • Did PayPal’s AI-Native Checkout Push with Perplexity Just Shift PayPal Holdings’ (PYPL) Investment Narrative?
  • AI Back-End Networks Continue Their Shift to Ethernet, Now Accounting for Over Two-Thirds of 3Q 2025 Switch Sales in AI Clusters, According to Dell’Oro Group
  • AMD's Lisa Su doesn't believe there's an AI bubble: 'Emphatically, from my perspective, no'
  • Meta enters AI licensing fray, striking deals with People Inc., USA Today Co. and more
  • Gemini 3 Pro: the frontier of vision AI