Stephen Hawking told Last Week Tonight’s John Oliver recounts a terrifying but remembered hypothetical story from a decade ago concerning the potential hazards of AI. The premise is that a group of scientists create a superintelligent computer and ask it, “Is there a God?” The computer responds, “There is now,” and a bolt of lightning strikes the plug, preventing it from being shut down. Let’s hope that’s not the case with OpenAI and some missing evidence in the New York Times’ plagiarism complaint.
According to Wired, a court declaration filed by the New York Times on Wednesday alleges that OpenAI’s programmers mistakenly removed evidence of the AI’s training data, which took a long time to investigate and collect. OpenAI recovered part of the data, but “the original file names and folder structure” that appeared when the AI put the articles into its models remain absent.
OpenAI spokesperson Jason Deutrom disagreed with the NYT’s assertions, stating that the business “will file our response soon.” The Times has been suing Microsoft and OpenAI since December of last year for alleged copyright infringement using its AI models.
The action is still in the discovery process, which is when both sides request and submit material to help develop their case for trial. OpenAI was required to pass over its training data to the Times, but it has yet to publicly release the specific material used to create the AI algorithms.
Instead, OpenAI set up a “sandbox” of two virtual machines for the NYT’s legal team to perform their investigation. Before deleting the material, the NYT’s legal staff spent over 150 hours combing through it on one of the machines. OpenAI admitted the deletion, but the company’s legal staff described it as a “glitch.” Although OpenAI engineers attempted to remedy the error, the recovered data lacked the NYT’s work. As a result, the New York Times had to almost start over. The NYT’s lawyers stated that they had no basis to suspect the deletion was purposeful.
