On September 5, the Seattle Times and Newsday filed a lawsuit against OpenAI and Microsoft in the New York federal court. Additionally, a coalition of nearly 400 U.S. news publishers filed a lawsuit against OpenAI and Microsoft. The plaintiffs allege widespread unauthorized data harvesting of copyrighted journalism, including paywalled archives, to train generative AI models like ChatGPT and Microsoft Copilot.
The Anatomy of the Scraping Allegations
The complaints lodged by the publishing coalition paint a systematic picture of unauthorized digital ingestion. According to court documents, the tech giants systematically bypassed paywalls and technical access restrictions, copying hundreds of thousands of investigative pieces, news articles, and original features directly onto their private servers without compensating creators or securing licensing agreements.
Alan Fisco, president and CEO of the Seattle Times, noted that his organization spends millions of dollars annually to produce original reporting. The lawsuit asserts that OpenAI and Microsoft stripped away author credits and copyright metadata during ingestion. This vector of data extraction makes it difficult for publishers to prove the illicit use of their contents.
Furthermore, Microsoft finds itself deeply implicated through its financial and technical partnership with OpenAI. Having invested over $13 billion in OpenAI, Microsoft integrated these scraping outputs directly into Microsoft Copilot and Bing’s conversational search features. The legal filings argue these tools synthesize outputs so close to original reporting that end users have little incentive to visit primary publisher domains or maintain digital subscriptions.
Defensive Maneuvers and the Fair Use Battleground
OpenAI representatives have consistently defended their data ingestion pipeline. Drew Pusateri, speaking for OpenAI, maintained that the company’s models promote innovation, rely on publicly available data, and fall under the legal doctrine of “fair use” within U.S. copyright law.
Microsoft expressed surprise at the legal action initiated by the Seattle Times and Newsday, while simultaneously emphasizing an appreciation for local journalism and a willingness to seek resolution.
This action builds upon mounting legal pressure, mirroring previous high-profile confrontations such as the December 2023 complaint filed by the New York Times. However, the sheer scale of a 400-publisher coalition fundamentally alters the litigation landscape, transforming isolated publisher disputes into a coordinated industry-wide challenge.
Changing Regulatory Currents and Government Interventions
The legal battle lines are further complicated by shifting federal stances. The DOJ argued that leveraging copyrighted works for AI training could qualify as fair use, framing broad AI development as a critical pillar for national security, economic prosperity, and scientific advancement.
Legal counsel for the publisher coalition, led by former New Jersey Attorney General Matthew J. Platkin, strongly rejects this interpretation. The coalition’s legal brief warns that unbridled data harvesting threatens to render local journalism economically unsustainable, effectively eroding the foundational reporting network upon which democratic accountability relies.
The plaintiffs are demanding punitive damages, the disgorgement of profits accrued through the unauthorized exploitation of their archives, and mandatory judicial orders forcing OpenAI and Microsoft to purge all scraped publisher data and destroy model weights trained on those specific copyrighted corpora. With parallel international disputes unfolding—such as the European Court of Justice examining Google’s data extraction practices ahead of a September 2026 ruling—the outcomes of these U.S. federal proceedings will establish global precedents for synthetic intelligence development and digital intellectual property rights.
