Google Employees Play Personal Tutor To Bard After AI-Chatbot's Recent Flop-Show: 'Learns Best By Example'

Alphabet Inc. GOOG GOOGL is reportedly improving the accuracy of its newest language model for dialogue application, Bard — and it’s doing so through the manual contribution of its employees. 

What Happened: Last week, Google showed off its chatbot Bard at a launch event. It highlighted how the chatbot could provide text summaries of search results to questions — but Bard gave an incorrect answer, leading to much public backlash and sparking a debate about the chatbot’s accuracy. 

See Also: Redditor’s Wife Inherited Long-Dated Google Call Options: What’s A Prudent Way To Handle The Position?

On Wednesday, Google’s vice president for search, Prabhakar Raghavan, sent an email to employees, asking them to help the company ensure its AI-powered chatbot gets answers right, reported CNBC.

The internal memo in question included a link to a do’s and don’ts page with detailed instructions on how Google employees could fix the responses as they tested Bard internally. 

“Bard learns best by example, so taking the time to rewrite a response thoughtfully will go a long way in helping us to improve the mode,” the document read.

Why It’s Important: Previously, Google CEO Sundar Pichai asked employees to spend at least two to four hours exploring and trying Bard. He acknowledged that taking Bard to a level where it can become a reliable platform “will be a long journey for everyone,” the report noted. 

Pichai is facing criticism and ridicule from employees and users for how Bard’s launch was handled. In posts on Google’s internal forum “Memegen,” employees described the troubled launch as “rushed,” “botched,” and “un-Googley.”

Check out more of Benzinga's Consumer Tech coverage by following this link.

Read Next: Google Fiber Rolls Out 5Gbps Plan In These US Cities — Stream, Game, Work All At Once For Just $125/Month