Why Google Employees Are Questioning Bard’s Helpfulness

Screenshot of Google employees’ Discord chat. Source: Bloomberg

Community members believe Bard cannot answer even basic questions. An X (Twitter) user wrote:

“Anyone who has used Bard would probably agree. It can’t even answer basic questions about how Google’s own apps work, eg Analytics, reCAPTCHA etc. Go test it.”

Google Product Manager Believes That Output Cannot Be Trusted

There were discussions about the reliability of the data generated by the chatbots. Dominik Rabiej, a product manager for Bard, says that LLMs are not at a stage wherein users cannot trust their outputs without independently verifying them. He said:

“My rule of thumb is not to trust LLM output unless I can independently verify it”

LLMs are AI models trained with a large set of data to generate human-like outputs. Chatbots such as ChatGPT and Bard use LLMs.

As the data is not completely reliable, Rabiej suggests that it is ideal to use Google Bard for brainstorming purposes instead of relying on it only for information. As a matter of fact, when a user first starts using Bard, it shows a message stating, “Bard is an experiment.” The message further reads:

“Bard will not always get it right Bard may give inaccurate or offensive responses. When in doubt, use the Google button to double-check Bard’s responses.”