Google denies using ChatGPT data to train Bard

by nativetechdoctor
2 minutes read

After the move to monopolize data for the Bing chatbot from Microsoft, the Google chatbot was queried about requisitioning data from rival ChatGPT to train its Bard AI.

According to The Information, Google is ambitious to change the chances of its AI chatbots by forcing the DeepMind division to help the Google Brain team defeat OpenAI with a new project called Gemini.

This report also revolves around Jacob Devlin, a former artificial intelligence engineer at Google. Before his resignation, he worked tirelessly to warn Alphabet CEO Sundar Pichai along with other executives that Bard’s machine learning models were being trained using ChatGPT. Devlin himself has had to use this data to train Bard before, and as he warns, this move by Google would violate OpenAI’s terms of service, which also explains why the answer Bard chatbot’s responses tend to be similar to the ChatGPT chatbot’s responses

However, the tech giant has staunchly denied the allegations that Bard was secretly using data from ChatGPT. In a conversation with The Verge, Google spokesman Chris Pappas said: “Bard is not trained by any data from ShareGPT nor ChatGPT”.

Recently, AI Google lost the trust of users before rival ChatGPT. Google tried to catch up with Microsoft with Bard, but the chatbot received a lot of criticism for its weird and confusing responses to context. Even after Bard was announced in mid-March 2023, when asked about the data source, the chatbot replied that it was trained from Google’s internal data, including Google Search, Gmail, and other applications. Soon after, Google had to clarify that the Bard training data had nothing to do with Gmail or private data from users and that this misleading answer was likely due to a bug in the test language model.

Related Posts

Leave a Comment

Adblock Detected

Please support us by disabling your AdBlocker extension from your browsers for our website.