Google execs call on employees to help improve Bard search tool

Google has acknowledged that its artificial intelligence search tool Bard is not always accurate in its responses to user queries.

The company is now calling on its employees to help fix the tool's shortcomings.

Prabhakar Raghavan, Google’s vice president for search, sent an email to staff asking for their help in making sure that the company's new ChatGPT competitor gets its answers right. The email reportedly contained a link to a guide that included instructions on how employees should fix responses when testing Bard internally. Staff members were encouraged to rewrite answers on topics they understood well.

According to the guide, "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 model."

Earlier this week, CEO Sundar Pichai asked employees to spend two to four hours of their time on Bard, acknowledging that "this will be a long journey for everyone, across the field".

Raghavan echoed this sentiment, saying that "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".

Google announced its conversation AI technology last week, but the move did not go as planned. A series of missteps around the announcement pushed the company's stock price down nearly 9 per cent. Employees criticised Pichai for the mishaps, describing the rollout internally as "rushed," "botched" and "comically short-sighted".

To address Bard's mistakes, Google is turning to human knowledge. At the top of the guide, Google provides guidance for what to consider "before teaching Bard". The company instructs employees to keep responses "polite, casual, and approachable," and to maintain an "unopinionated, neutral tone".

Employees are told not to stereotype and to "avoid making presumptions based on race, nationality, gender, age, religion, sexual orientation, political ideology, location, or similar categories". The guide also instructs employees to avoid describing Bard as a person, implying emotion, or claiming to have human-like experiences.

Google encourages employees to give a "thumbs down" to answers that offer "legal, medical, financial advice," or are hateful and abusive. However, the company instructs employees not to try to rewrite the answer, as the Google team will take it from there.

To incentivize employees to test Bard and provide feedback, Raghavan said that contributors would earn a "Moma badge" that appears on internal employee profiles. Google will also invite the top 10 rewrite contributors from the Knowledge and Information organisation to a listening session. There, they can "share their feedback live" with Raghavan and the team working on Bard.

More from Business

News