Google denies Bard was trained with ChatGPT data (original) (raw)

Google’s Bard hasn’t exactly had an impressive debut — and The Information is reporting that the company is so interested in changing the fortunes of its AI chatbots, it’s forcing its DeepMind division to help the Google Brain team beat OpenAI with a new initiative called Gemini. _The Information_’s report also contains the potentially staggering thirdhand allegation that Google stooped so low as to train Bard using data from OpenAI’s ChatGPT, scraped from a website called ShareGPT. A former Google AI researcher reportedly spoke out against using that data, according to the publication.

But Google is firmly and clearly denying the data is used: “Bard is not trained on any data from ShareGPT or ChatGPT,” spokesperson Chris Pappas tells The Verge.

Pappas declined to answer whether Google had ever used ChatGPT data to train Bard in the past. “Unfortunately, all I can share is our statement from yesterday,” he says.

According to _The Information_’s reporting, a Google AI engineer named Jacob Devlin left Google to immediately join its rival OpenAI after attempting to warn Google not to use that ChatGPT data because it would violate OpenAI’s terms of service, and that its answers would look too similar. One source told the publication that Google stopped using that data after his warnings. Perhaps it threw out that portion of the training, too.

Regardless of whether Google did or didn’t wind up using the data, it’s interesting to hear that Google might be pulling in DeepMind, which has been trying and failing to become more independent from Google for many years.

Update March 30th, 2:02PM ET: Google would not answer a follow-up question about whether it had previously used ChatGPT data form Bard, only that Bard “isn’t trained on data from ChatGPT or ShareGPT.”