It has been reported that a Google employee left the company after warning CEO Sundar Pichai that using OpenAI’s ChatGPT to train their AI search chatbot Bard was a mistake.
The public has shared their conversations with ChatGPT on ShareGPT, but OpenAI prohibits the use of their outputs to train other models.
Jacob Devlin, an AI engineer at Google, was concerned that using this data would violate OpenAI’s terms of service and make Bard too similar to ChatGPT.
After he raised his concerns to Pichai, he reportedly quit and joined OpenAI. The competition to develop and deploy the most impressive generative AI products among tech giants, particularly between Google and Microsoft’s OpenAI, is intense.
While it may be challenging to avoid using text generated by competing models in the future, intentionally doing so would be deemed inappropriate.
A spokesperson from Google informed The Verge that the company refuted any use of text produced by ChatGPT to train Bard.
The spokesperson explicitly stated, “Bard is not trained on any data from ShareGPT or ChatGPT.” Nonetheless, the representative did not provide any commentary regarding whether Google had ever employed ChatGPT-generated text to train Bard.
