Skip to content

Instantly share code, notes, and snippets.

@satish860

satish860/openaivsnyt.md

Created May 23, 2024 05:15
Show Gist options
  • Save satish860/5202fdb950ec738a6a388c7f4d969ac6 to your computer and use it in GitHub Desktop.
Save satish860/5202fdb950ec738a6a388c7f4d969ac6 to your computer and use it in GitHub Desktop.
Download ZIP
Raw
openaivsnyt.md

Timeline of Events in The New York Times Company v. OpenAI and Microsoft

1851

  • The New York Times was founded, establishing itself as a long-standing, reputable news organization. (Complaint, page 8)

December 2015

  • OpenAI was formed as a non-profit artificial intelligence research company, marking the origins of the defendant company. (Complaint, page 16)

March 2019

  • OpenAI created OpenAI LP, a for-profit company, transitioning into a commercial entity and setting the stage for alleged infringement. (Complaint, page 17)

2019

  • Microsoft invested $1 billion in OpenAI, demonstrating the beginning of the partnership between the two companies. (Complaint, page 18)
    • Evidence: The complaint details the financial and technological partnership between OpenAI and Microsoft, including Microsoft's substantial investments. (Complaint, pages 15-19, 56-58)

2020

  • OpenAI released GPT-3, a large language model (LLM) with 175 billion parameters trained on vast amounts of text data to predict the next word in a sequence. This release is key to the allegations of copyright infringement and misappropriation. OpenAI also ended its commitment to openness. (Complaint, pages 18, 26)

November 2022

  • OpenAI released ChatGPT, a chatbot based on GPT-3.5, which gained over 100 million users within three months. The rapid growth of ChatGPT's user base is relevant to the scope of alleged infringement and damages. (Complaint, page 18)

February 2023

  • Microsoft unveiled Bing Chat, powered by GPT-4, an even larger LLM developed by OpenAI with alleged capabilities surpassing prior models. The integration of GPT-4 into Microsoft's products supports claims of joint infringement and commercial benefit. (Complaint, pages 18, 27)
    • Evidence: The integration of OpenAI's AI technology into Microsoft's products, like Bing Chat, is used to argue for the defendants' joint liability and commercial benefits derived from the alleged infringement. (Complaint, pages 15-19, 56-58)

May 2023

  • Microsoft and OpenAI unveiled "Browse with Bing," a plugin allowing ChatGPT to access current web content, enabled by a technique called "retrieval augmented generation." This further demonstrates the integration of OpenAI's technology into Microsoft's products, strengthening claims of partnership. (Complaint, pages 18, 23)

As of December 31, 2022

  • Plaintiff The New York Times Company employed approximately 5,800 full-time employees, highlighting the scale of its operations and resources dedicated to producing original content. (Complaint, page 12)
    • Evidence: The Times's efforts to protect its intellectual property through paywalls, terms of service, and licensing agreements support the argument that the defendants' use of its content was unauthorized. (Complaint, pages 13-16)

2023

  • Microsoft increased its investment in OpenAI to a reported $13 billion, deepening the financial ties between the two companies and the potential for commercial benefit from the alleged infringement. (Complaint, page 6)
    • Evidence: Microsoft's significant investment in OpenAI is used to argue for the defendants' joint liability and highlight the commercial benefits they derived from the alleged infringement. (Complaint, pages 15-19, 56-58)

April 2023

  • The Times reached out to Defendants regarding intellectual property concerns, putting them on notice of potential infringement. (Complaint, page 15)

By the third quarter of 2023

  • The Times had nearly 10.1 million digital and print subscribers, demonstrating the widespread reach and potential impact of the alleged infringement on its business. (Complaint, page 13)
    • Evidence: The Times's large subscriber base highlights the potential impact of the alleged infringement on its business and the importance of protecting its intellectual property. (Complaint, pages 13-16)

April 2023

  • ChatGPT had approximately 173 million users, further illustrating the significant scale of the alleged infringement and its potential to harm The Times's market position. (Complaint, page 57)

As of August 2023

  • OpenAI was on pace to generate more than $1 billion in revenue over the next twelve months, highlighting the substantial commercial benefits derived from the AI products that allegedly infringe on The Times's copyrights. (Complaint, page 58)
    • Evidence: OpenAI's significant revenue is used to demonstrate the commercial benefits gained from the alleged infringement of The Times's copyrights. (Complaint, pages 56-58)

Date Unspecified

  • Defendants allegedly trained their AI systems using Times content without authorization. Training AI models involves feeding them large datasets, including scraped web content, to "learn" patterns and representations. This forms the core allegation supporting claims of direct copyright infringement and violation of terms of service. (Complaint, pages 27-29)

    • Evidence: The Times's efforts to protect its intellectual property through paywalls, terms of service, and licensing agreements support the argument that the defendants' use of its content for training AI models was unauthorized. (Complaint, pages 13-16)

  • Defendants' AI products allegedly reproduce verbatim copies of Times articles, an example of "memorization" where language models reproduce snippets of their training data. This supports claims of copyright infringement through unauthorized reproduction and distribution of Times content. (Complaint, pages 29-32)

    • Evidence: Exhibit J of the complaint provides multiple examples where OpenAI's GPT-4 model reproduces significant portions of Times articles verbatim when prompted, supporting the claim of direct copyright infringement. (Complaint, pages 29-32, Exhibit J)

  • Defendants' synthetic search results allegedly include excerpts and paraphrases of Times articles, generated through "retrieval augmented generation" where LLMs are prompted with web search results. This is relevant to claims of infringement and misappropriation of Times content for commercial purposes, as it diverts potential readers from The Times's own properties. (Complaint, pages 38-47)

    • Evidence: The complaint includes examples of Bing Chat and ChatGPT's Browse with Bing feature generating search results that include extensive excerpts and paraphrases of Times articles, supporting claims of copyright infringement and misappropriation of content. (Complaint, pages 38-47)

  • Defendants' AI products have allegedly generated false information attributed to The Times, a phenomenon known as "hallucination" in AI where models confidently produce incorrect outputs. This supports claims of unfair competition and trademark dilution through false attribution and tarnishment of The Times's reputation for accuracy and reliability. (Complaint, pages 52-55)

    • Evidence: Several instances are provided where the defendants' AI products generate and attribute false information to The Times, such as fabricated article titles, quotes, and recommendations, supporting claims of unfair competition and trademark dilution. (Complaint, pages 52-55)
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment