Reportedly, Google is preparing to launch Bard for the use of the general public. As the company instructs its employees to help in training the AI by rewriting its bad or wrong answers.
As reported by CNBC that Google’s VP of Search Prabhakar Raghavan, sent an email to employees ahead of this week. This is intended to ask for help in order to train Bard by correcting incorrect or bad responses. Since this email also comprises a page of do’s and don’ts for making understand their employees how to correct Bard. It also incentivized employees to write the corrections with a “Moma badge,” the same going to appear on internal employee profiles.
Additionally, the do’s and don’t document requests the staffers to interact with Bard in order to help test its capacity. It also speeds up the training of the model. Whereas, on the topics that can be “understand well” by employees.” The document requests the employees to rewrite the answers on their own that are incorrect.
“The document says in part: 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]”
Further, it goes on to add that when employees rewrite answers then they should do so in a “polite, casual and approachable” manner. At the same time, keeping an “unopinionated, neutral tone,” and writing in the first person. In addition to this, the document further states that responses should not stereotype and “avoid making presumptions based on race, nationality, gender, age, religion, sexual orientation, political ideology, location, or similar categories.” Whereas, Answers also shouldn’t “imply emotion” or “claim to have human-like experiences.”
Also, when it comes to more sensitive topics, such as “legal, medical, [or] financial advice” or in case if Bard responds with something hateful or abusive. The employees are told to give a “thumbs down.”
As per the email, the “top 10” contributors from the Knowledge and Information organization would also get the chance to share feedback directly with Raghavan.
“Raghavan said: This is exciting technology but still in its early days. We feel a great responsibility to get it right, and your participation in the dog food will help accelerate the model’s training and test its load capacity (Not to mention, trying out Bard is actually quite fun!).”
Moving ahead with this news, it comes as various have found in the past week that Microsoft’s Bing AI has a tendency to go off the rails after a certain point. On the other hand, it starts making responses that go directly against things that Google certainly trying to avoid in this training.