OpenAI Remove ChatGPT Training Data, Evidence Of Violation Of Missing Copyright

JAKARTA OpenAI has made a big mistake that may harm its company. Evidence regarding ChatGPT training using copyrighted data has been accidentally deleted.

Some time ago, OpenAI was sued by two major media, namely the New York Times and Daily News, for using their work without permission to train Artificial Intelligence (AI)-based chatbots. To support their lawsuit, they conducted research.

Lawyers from these two media conducted a search into the ChatGPT training data set using two virtual machines. In searching for evidence, the New York Times and Daily News even hired experts.

They have spent more than 150 hours since November 1 to train OpenAI research data. However, this search was hampered, and even had to stop, because the data was deleted on November 14.

According to a TechCrunch report, technicians at OpenAI removed all search data examined by both media in one of the virtual machines. That way, the search for evidence becomes in vain because the data has been deleted by the company.

The two media filed their complaint with the US District Court for the South District of New York on November 20. Shortly after his complaint was made public, OpenAI admitted his guilt.

The ChatGPT developer said that they deleted research data of the two media companies by accident. Some of the data was saved, but incomplete and unusable.

On the other hand, the New York Times does not believe that OpenAI's actions were done by accident. According to him, there is no reason to believe that the deletion of this research data was done accidentally.