OpenAI and Microsoft Aware of the Web’s Impending ‘Doom Loop’
Recently revealed court documents in the case against OpenAI and Microsoft paint a troubling picture. Internal communications from the companies warned of a “doom loop” that could harm the web, described their data scraping practices as the “largest theft of labor in human history,” and claimed they made a “complete mockery of the idea of fair use.”
A lot of the striking quotes come from Brent Hecht, Microsoft’s Director of Applied Science, although the company has attempted to distance itself from his comments. A Microsoft spokesperson stated that “These comments reflect one employee’s individual perspective, are not a legal analysis, and do not represent the company’s views.”
In another court filing, Jordan Usdan, GM for Data Strategy and Ops at Microsoft AI, characterized Hecht’s views as adversarial, noting that Hecht “holds divergent, academic, and forward-looking views about how data ecosystems for AI should operate” and does not speak for Microsoft regarding his theoretical perspectives on AI’s impact on content creators.
Regardless of Microsoft’s attempts to distance itself, the implications of these statements are undeniable. The concept of Google Zero is becoming a reality, and AI is indeed reshaping the web.
The court filing includes numerous bold statements from various figures, including Satya Nadella and Sam Altman, along with other OpenAI employees. Here are some key takeaways from the extensive document.
“An astonishing theft”
The introduction highlights comments from Hecht and OpenAI’s Head of ChatGPT, indicating that the companies recognized they posed an “existential threat” to publishers like the New York Times. Hecht referred to the data harvesting by ChatGPT and Copilot as the “largest theft of labor in human history,” asserting that Microsoft’s defense undermines the concept of “fair use.”
It’s a “doom loop”
Satya Nadella acknowledged that chatbots have effectively replaced traditional search methods, eliminating the need to consult original sources for information. An internal Microsoft document further states, “Our AI content strategy has started a ‘doom loop’ that will hurt the performance of our models and the entire web at the same time.” It’s unusual for a product to threaten the economic foundations of its essential suppliers, yet that’s the situation created for their LLM business regarding its “content supply chain.”
That’s not even a real number
Don’t be misled by OpenAI or Microsoft’s claims of altruism. OpenAI cofounder Greg Brockman appears more focused on the potential “gazillions” of dollars to be made through commercial AI.
Paywall shmaywall
Despite Nadella’s later assertion that “anything that is paywalled should be licensed,” an OpenAI representative admitted to being “unaware” of any efforts to filter out paywalled content from training data.
“Insanely good at regurgitation”
Internally, OpenAI seemed to recognize ChatGPT’s tendency to reproduce copyrighted material “verbatim.” While they acknowledged the importance of preventing memorization to minimize copyright violations, employees conceded that GPT-4 “memorized a ton of data and therefore will be insanely good at regurgitation.” The filing cites instances of ChatGPT producing lengthy excerpts from articles in various publications in response to user queries.
“‘Hoovering up’ all their work”
Microsoft was aware of how its extensive data scraping would be perceived, admitting that “almost no one intended for the content they created to be used in this fashion, nor are they compensated for its use.”
A “substitute for the labor of people”
OpenAI Policy Director Jack Clark recognized the implications, stating that they were “creating systems that substitute for the labor of the people that define the ‘culture’ of society.” Internal documents referred to ChatGPT as “the modern newsstand,” with OpenAI’s Nick Turley later commenting that once users receive an answer from the chatbot, there is “no good reason to click” on a link to the source.
Destroying their own supply chain
Microsoft acknowledged that “LLMs are a product that destroys its own supply chain” because they often serve as substitutes for their own training data.
OpenAI knows it’s killing referral traffic
OpenAI’s media and economic experts linked the decline in referral traffic for sites like the Times directly to AI-generated summaries, speculating that search referrals may have dropped by as much as 60 percent.
A Microsoft spokesperson emphasized that “Satya’s testimony and Microsoft’s position in this case are perfectly consistent,” clarifying that his comments addressed broad principles and changes in how people find and consume information, which should not be conflated with copyright issues before the Court.
However, it’s evident from these newly unsealed documents that both Microsoft and OpenAI were aware they would cause significant harm to the publishing industry and the millions it employs, all in pursuit of “gazillions” of dollars—doom loop consequences notwithstanding.
