The way staff members are testing Google‘s Bard chatbot serves as a prime example of the company’s long-standing commitment to open discourse and debate. The AI chatbot’s conversational abilities might be improved by asking it questions and reporting poor responses.

As chatbots such as OpenAI’s ChatGPT and Bard learn to write by analyzing human-written material, according to a company-wide email from CEO Sundar Pichai last week. Bard will be incorporated into search and capable of conversing with users about a range of subjects, relying on indexed webpages for up-to-date knowledge.

Employees are already grumbling that Pichai’s move comes just a short time after Google abruptly began laying off about 12,000 workers in response to sluggish revenue growth after a flu surge. Employees posted memes on internal message boards and urged Bard to comment on recent layoffs and employee treatment in talks they had with the company. These memes and queries imply that staff members aren’t very thrilled about getting more work at a time when people feel their work is being evaluated more closely. In one of these exchanges, which Insider has reviewed, a coworker asked Bard to envision an IT firm that had laid off 12,000 workers.

Internally, Google has already received criticism over the Bard announcement. Google’s unveiling of the chatbot appeared hurried because Microsoft recently released a redesigned version of Bing that incorporates ChatGPT. Moreover, advertising for Bard showed a chatbot answering a question wrong, and the following day, the company’s stock fell by almost 9%.

Google has stated that it intends Bard to offer accurate, useful responses to inquiries that may not have a single correct response and then refer visitors somewhere to delve deeper. The threat posed by an inaccurate answer is substantial, thus staff members have been advised to report any responses provided by Bard that appear to offer financial or medical advice. Also, the bot shouldn’t sound overly human. Google asked staff members for assistance because doing so necessitates training the bot to steer clear of such topics.

Along with worries that Bard might unintentionally say anything debatable, Google’s search revenue may be negatively affected if consumers rely on Bard to answer their queries rather than clicking promoted links. However, using an AI chatbot for searches costs extra. Given that ChatGPT has been used by more than 100 million people since its release, Google probably doesn’t want to take a chance on losing control of search to Microsoft. An inquiry for comments was not immediately answered by Google.