Google executives provided employees with a guide containing the do’s & don’ts to rectify any unfavorable outcomes of the Bard AI tool
Staffers were instructed to remain “neutral” and “not imply emotion”
Google Executives are aware that their AI search engine, Bard, isn’t always spot-on in its responses to user queries. The responsibility is being placed on employees to correct any incorrect answers given by Bard.
Prabhakar Raghavan, Google’s vice president for search, sent out an email to his employees on Wednesday. He asked them to help ensure that the company’s ChatGPT rival provides accurate responses. CNBC had access to this message which included a link with guidelines and do’s & don’ts for staff members as they review Bard internally.
Staffers are urged to craft answers for topics they have an extensive knowledge about.
The document reads, “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.”
As CNBC reported earlier, CEO Sundar Pichai asked employees on Wednesday to allot two to four hours of their day toward the Bard initiative. He acknowledged, “this will be a long journey for everyone, across the field.”
Raghavan mirrored that opinion. He wrote, “This is exciting technology but still in its early days. We feel a great responsibility to get it right, and your participation in the dogfood will help accelerate the model’s training and test its load capacity (Not to mention, trying out Bard is actually quite fun!).”
Last week, Google revealed its conversation technology. But the missteps that occurred during the announcement caused their stock price to plummet nearly 9%. Employees were vocal about CEO Pichai’s management of the launch. They described it internally as “rushed,” “botched” and “comically short sighted.”
In an effort to rectify AI mistakes, the executives are turning to the knowledge and wisdom of humans. Atop its list of dos and don’ts, Google has published directions for proper guidance for what to consider “before teaching Bard.”
Under do’s Google asked employees to keep responses “polite, casual and approachable.” It also suggested that they should be “in first person,” and maintain an “unopinionated, neutral tone.”
For don’ts, Google instructed employees not to stereotype and to “avoid making presumptions based on race, nationality, gender, age, religion, sexual orientation, political ideology, location, or similar categories.”
The document says, “don’t describe Bard as a person, imply emotion, or claim to have human-like experiences.”
Google asks to “keep it safe,” and to give a “thumbs down” to answers that show “legal, medical, financial advice” or are hateful and abusive.
It also says, “Don’t try to re-write it; our team will take it from there.”
