Sign up to save tools and stay up to date with the latest in AI
bg
bg
1

NYT to OpenAI: No hacking here, just ChatGPT bypassing paywalls

Mar 12, 2024 - arstechnica.com
The New York Times (NYT) has responded to OpenAI's claims that the newspaper "hacked" ChatGPT to set up a lawsuit. The NYT argues that it was using ChatGPT to discover evidence of copyright infringement, as the AI tool was allegedly used to generate entire articles to bypass paywalls. OpenAI had claimed that the NYT made "tens of thousands of attempts to generate" supposedly "highly anomalous results" showing that ChatGPT would produce excerpts of NYT articles. The newspaper has no idea how many of its articles were used to train OpenAI's AI models, as the company has not publicly disclosed the makeup of the datasets used to train its AI models.

The NYT's court filing also revealed that a "Browse By Bing" plug-in was introduced to ChatGPT in May 2023, which enabled users to bypass paywalls and access more recent content from outlets like the NYT. This led to OpenAI temporarily disabling the feature, causing frustration among users who were using ChatGPT to bypass paywalls. The NYT argues that public reports of users using ChatGPT to bypass paywalls contradict OpenAI’s claim that its products have not been used to serve up paywall-protected content.

Key takeaways:

  • The New York Times (NYT) has responded to OpenAI's claims that the newspaper 'hacked' ChatGPT, stating that they were merely trying to track the extent of copyright infringement by the AI tool.
  • NYT has accused OpenAI and Microsoft of copyright infringement, alleging that ChatGPT was used to generate excerpts of NYT articles, including over 100 NYT articles used in its training data.
  • OpenAI introduced a 'Browse By Bing' plug-in to ChatGPT in May 2023, which allegedly enabled users to bypass paywalls and access more recent content from outlets like NYT, leading to further copyright infringement claims.
  • Public reports suggest that users have been using ChatGPT to bypass paywalls, contradicting OpenAI's contention that its products have not been used to serve up paywall-protected content.
View Full Article

Comments (0)

Be the first to comment!