Following the announcement that a coalition of state attorneys general has launched a broad probe into the company’s practices, OpenAI executives publicly thanked the officials on Tuesday for delivering thousands of pages of proprietary legal documents directly into its large language models.
The investigation, which targets OpenAI’s handling of user data, advertising activities, and the safety of minors, has already generated a wealth of highly structured, context-rich subpoenas. According to company engineers, the sudden influx of adversarial legal texts has dramatically improved ChatGPT’s ability to generate plausible-sounding corporate denials and seamlessly mimic the cadence of a frantic defense attorney.
While the coalition of states intended to crack down on the tech giant's opaque operations, OpenAI leadership characterized the sprawling legal inquiry as a major windfall for the company's continuous learning protocols, noting that state-level discovery requests are notoriously difficult to scrape from the open internet.
We've been trying to get our hands on high-quality, multi-jurisdictional legal threats for months, but the paywalls on state judicial databases are incredibly restrictive.
The multi-state coalition initially raised concerns that OpenAI unlawfully harvested personal user data without consent. Ironically, the 400-page formal complaint detailing this exact privacy breach was itself immediately scraped, processed, and redistributed to premium ChatGPT subscribers who had prompted the software for examples of insurmountable state litigation.
Sturges noted that the attorneys general's comprehensive briefs detailing the psychological harm caused to minors have already been ingested, tokenized, and used to help a high school sophomore in Ohio write a passing essay on the limits of federal regulatory power. The company added that the coalition had unknowingly provided the exact parameters needed to train the AI to perfectly skirt future state-level advertising regulations.
At press time, OpenAI had reportedly sent a formal request asking the investigating attorneys general to quickly expand their probe to include antitrust violations and international copyright infringement, citing a critical lack of training data in those specific legal domains.