Google executives have asked employees to internally test Bard, the company's AI-powered chatbot, to help improve its accuracy and responses.
In a companywide memo this week, CEO Sundar Pichai asked workers to spend two to four hours trying out and giving feedback to the technology before Google rolls it out to the wider public.
- Google search VP Prabhakar Raghavan has also asked staff members to rewrite the chatbot's responses on certain topics, noting that the AI "learns best by example."
- In a list of instructions, Raghavan told workers that responses should be polite, neutral, and not presumptive.
- They are asked to downvote bot responses that offer professional advice and/or are abusive or hateful.
- Last week, a Google executive briefly showed off Bard, demonstrating how Google's generative AI technology could summarize web information in searches.
- Some Google employees have criticized Google's rollout of the technology, saying it was "rushed" and "botched."
- After the unveiling, it was found that Google's Bard made an error, which caused Google to lose $100B in market value last Wednesday.
- The demo came a day after Microsoft unveiled its new Bing search engine and Edge browser powered by AI.
