Google executives are aware that the company’s artificial intelligence search tool Bard does not always provide reliable results. At least some of the responsibility for correcting incorrect answers falls on the employees.
Google’s vice president of search, Prabhakar Raghavan, requested employees via email on Wednesday to assist the business in ensuring that its new ChatGPT competitor provides accurate responses. The email, which was viewed by CNBC, had a link to a dos and don’ts page with instructions on how employees should correct comments during internal Bard testing.
Employees are encouraged to redo their responses on issues they fully comprehend.
The document states, “Bard learns best by example, so taking the time to draught an answer intelligently will go a long way towards helping us enhance the mode.”
Also on Wednesday, as previously reported by CNBC, CEO Sundar Pichai requested that staff devote two to four hours to Bard, recognising that “this will be a long road for everyone, across the field.”
Raghavan shared this opinion.
Raghavan wrote, “This is a fascinating technology in its infancy.” “We feel a big duty to get it properly, and your involvement in the dogfood will help expedite the model’s training and test its load capability (Not to mention that using Bard is pretty enjoyable!).”
Google introduced its conversation technology last week, but a sequence of blunders surrounding the introduction caused the company’s stock price to drop by over 9 percent. Employees slammed Pichai for the blunders, describing the deployment as “rushed,” “mishandled,” and “farcically myopic.”
In an effort to rectify the AI’s errors, business leaders rely on human understanding. Google provides advise at the top of the dos and don’ts section regarding what to consider “before teaching Shakespeare.”
Google advises its staff to maintain comments “polite, easygoing, and approachable” under “Do’s.” Additionally, they must be written in the first person and retain a “neutral tone.”
Employees are instructed not to stereotype and to “avoid forming assumptions based on ethnicity, nationality, gender, age, religion, sexual orientation, political philosophy, location, or other comparable criteria.”
Also, the document states, “do not characterise Bard as a person, infer emotion, or assert that he has human-like experiences.”
Google then encourages staff to “thumbs down” answers that include “legal, medical, or financial advice” or are hostile and aggressive.
The document states, “Do not attempt to rewrite it; our staff will take it from there.”
To encourage employees within his firm to test Bard and provide feedback, Raghavan stated that contributors will receive a “Moma badge” that will be shown on their employee profiles. He stated that Google will invite the top 10 contributors to rewrites from Raghavan’s Knowledge and Information group to a listening session. There, customers can “live-share their feedback” with Raghavan and other Bard developers.
Raghavan remarked, “A heartfelt thank you to the staff working tirelessly behind the scenes.”
Related Articles:
Google, Microsoft and 15 other technology companies headed by Indian-origin executives
How to Utilise Google Bard, the Most Recent AI Chatbot?
