Newly unsealed court documents show that executives at Microsoft and OpenAI had significant concerns about how their AI systems were built using publishers' content.

According to The Hollywood Reporter, a 2023 internal Microsoft document described the ingestion of copyrighted works as "an astonishing theft of unprecedented proportions" and potentially "the largest theft of labor in human history." Microsoft Director of Applied Science Brent Hecht is cited using that language.

The filings come in a lawsuit brought by The New York Times and eleven other publishers against Microsoft and OpenAI over the use of their articles to train AI systems. The publishers call it theft; the companies say the secret to their technology was innovation.

The documents describe an exchange in which OpenAI president Greg Brockman was told of a "hack to get around [the] nytimes paywall" and replied, "Ah nice." Publishers argue employees were instructed to devise ways to circumvent paywalls "without detection." Microsoft CEO Satya Nadella stated that "anything that is paywalled should be licensed by anyone who wants to use it," and said he would have invoked Microsoft's right to require retraining if he had known restricted content was scraped.

At this stage, OpenAI and Microsoft bear the burden of proving fair use. Publishers point to examples of near-verbatim regurgitation, including Bing Chat copying all but two of the first 396 words of a 2023 Times article. Internal Microsoft data reportedly shows click-through rates falling 87% to 93% when users searched through Bing Chat rather than traditional search. Hecht said a ruling for the companies would "make a complete mockery of the idea of fair use."

Source: The Hollywood Reporter