Human personnel are used by Google to enhance Its chatbot answers

Google’s yet-to-be-launched AI chat bot Bard boldly spouted disinformation about the James Webb Space Telescope in a video commercial released on Twitter. The chatbot said, “JWST took the very first photographs of a planet outside of our own solar system,” which is clearly incorrect. (For the first time, photos of exoplanets were acquired by the European Southern Observatory’s Very Large Telescope.) According to CNBC, the software titan is now attempting to increase Bard’s accuracy and has asked staff for assistance.

Prabhakar Raghavan, Google’s VP of Search, allegedly issued an email to staff workers requesting them to rewrite Bard replies on areas they are familiar with. According to Raghavan, the chatbot “learns best by example,” and training it with factual replies would help enhance its accuracy. According to the email obtained by CNBC, Raghavan also offered a list of “dos” and “don’ts” for correcting Bard’s reply.

Answers should be written in the first person, be unopinionated and impartial, and have a courteous, informal, and accessible tone. Workers are also told not to “make assumptions based on ethnicity, nationality, gender, age, religion, sexual orientation, political philosophy, geography, or other comparable criteria.” They are forbidden from referring to Bard as a person, implying that it has feelings, or claiming that it has human-like experiences. Moreover, users are encouraged to thumbs down any comments from the chatbot that provide “legal, medical, or financial advice” or are hostile and abusive.

Raghavan’s statement came after Google CEO Sundar Pichai asked staff to spend a few hours each week testing the AI chatbot in an email. Employees at Google allegedly chastised Pichai for a “hurried” and “botched” Bard deployment. Employees may now “help shape [the chatbot] and participate” by evaluating the company’s new product, according to the CEO. He also reminded everyone that some of Google’s “most successful products” were not “first to market,” but “gained momentum because they fulfilled critical customer requirements and were founded on profound technological insights.”

With the arrival of the OpenAI chatbot late last year, many have been waiting for Google’s reaction to ChatGPT. The Microsoft-backed technology has grown in popularity in recent months, enough to frighten Google and its investors. Google sought to ease investors’ fears during its quarterly results call in early February by talking about its own chatbot and by touching on its work on an AI-powered Search to compete with the next-gen Bing.