Tech News Desk – Google is working hard on its new AI tool Bard to compete with Chat GPT. Currently, Bard has been made live for some users. Meanwhile, a shocking news is coming out about Google’s AI tool. Actually, Bard has claimed that he has been trained on the basis of people’s Gmail data. This controversial answer of Bard has been shared by Microsoft researcher Kate Carford via Twitter.
When Kate asked Bard what the chatbot’s dataset was, Google’s AI tool replied that its data was drawn from a variety of sources, including Wikipedia, GitHub, Stack Overflow, and Gmail. It became clear from this answer that Google’s new AI tool has been trained on the basis of Gmail data. If so, then this is the biggest mess with people’s privacy.
Responding to Kate’s tweet in this regard, Google said that Bard is still in the development phase, trained on large language models and may make mistakes in the beginning. Google also clarified that its data was not taken from Gmail. Be warned, AI tools like Bard and Chat GPT won’t always give you the right answer. Sometimes there may be factual mistake in these. Google itself has informed people through a blogpost that the AI tool can sometimes give wrong information.
Bard is currently live for select users in the UK and US. Its testing is going on and other people have been kept in the waiting list. At the same time, a new version of Chat GPT has been launched by GPT-4 Open AI. The new version is more advanced and accurate than before. In this, users can also ask questions through photos. Access to GPT-4 is only available to those who have purchased the subscription model of Chat GPT.