Google parent company Alphabet is warning employees not to enter confidential materials into chatbots, including its own chatbot Bard, Reuters reported Thursday, joining a growing list of companies concerned about sensitive internal information being leaked through AI.
Google warned Alphabet engineers Thursday to avoid direct use of computer code that chatbots can also produce, because AI can reproduce the data it absorbs during training, risking a potential leak, Reuters reported, citing “four people familiar with the matter.”
Potential leaks from the AI technology could help Bard’s competitor ChatGPT in the ongoing race to dominate the AI world where billions of dollars in investment and advertising are still up for grabs.
Early this week, Politico reported Google was postponing Bard’s launch in the European Union because of concerns about the chatbot’s impact on privacy.