According to the inside of The Information, Google has encountered problems with its Bard chatbot and is now turning for help to its DeepMind division to beat rival OpenAI with a new initiative called Gemini.
The post also suggests that Google may have used ChatGPT data to train Bard. They were allegedly taken from ShareGPT, a website that analyzes conversations conducted using OpenAI chat models.
Using the data would violate OpenAI’s terms of service, and former Google AI researcher Jacob Devlin is said to have left Google to join OpenAI after warning the company not to use ChatGPT data. Devlin was concerned that Google was violating OpenAI’s terms of service and that Bard’s responses would be too similar. According to one source, Google stopped using the data after Devlin’s warnings.
Google denied the allegations, and spokesman Chris Pappas told The Verge that Bard was not trained on data from ShareGPT or ChatGPT. All that said, it’s interesting to hear that Google is reaching out to DeepMind, a company that has been trying to become more independent of Google for years, but so far to no avail.
As Google tries to catch up with Microsoft, use of AI chatbots raises concerns about potential job losses for humans: Goldman Sachs analysis suggests that AI could result in the loss of up to 300 million jobs.
SpaceX CEO Elon Musk, Apple co-founder Steve Wozniak and other famous people even called for a freeze on AI research so government regulators could catch up.