In a landmark case filed in December 2023, The New York Times initiated legal proceedings against OpenAI and Microsoft, alleging unauthorized use of its copyrighted content. The litigation centers on accusations that OpenAI's ChatGPT and GPT-4 models were trained on articles from The New York Times without proper authorization, a claim bolstered by the news organization's presentation of 100 instances in support of its argument.
OpenAI's Retort Suggests Times Hack
OpenAI has responded to these allegations by suggesting The New York Times engaged in questionable tactics to prove its point. According to OpenAI, The Times orchestrated a series of tens of thousands of attempts to extract responses from ChatGPT by exploiting a specific bug, in violation of OpenAI's terms of use. These attempts were aimed at generating “highly anomalous results,” demonstrating the capability of the AI to regurgitate verbatim passages from Times' articles. OpenAI has pledged to address the exploited bug and emphasizes its commitment to supporting journalism without monopolizing factual information or language rules.
The Microsoft Angle and Looking Ahead
The lawsuit extends beyond OpenAI to implicate Microsoft through its Bing Chat/Copilot service, which integrates ChatGPT technology to potentially generate content resembling that of Wirecutter, The New York Times' product review site. The Times argued this led to a decrease in referral-based revenue. However, OpenAI counters this claim, suggesting that the evidence presented was the result of specific prompts by The Times designed to elicit non-verbatim summaries directing users back to the Wirecutter site.
As the legal battle unfolds, the disagreement highlights significant questions about AI, copyright, and the evolving relationship between technology companies and content creators. At this stage, official comments from the involved parties regarding the latest court filings are pending, leaving industry observers awaiting further developments in this precedent-setting case.
