The developers complained that their lives had become hell after Bard entered the competitive race with ChatGPT: the workload and complexity of tasks increased exponentially, and they had to do things they absolutely did not understand.

Google wants to win at any cost

According to the source, in the process of training the chatbot and verifying the information it provides to users, Bard team members are often assigned tasks with no specific expertise. For example, they evaluated Bard’s responses on state laws, medicine, drugs, and more. Moreover, all this had to be done in a very short time, usually no more than three minutes.

Right now people are scared, stressed, underpaid, don’t know what’s going on, and this culture of fear isn’t helping you get the quality and teamwork you want from us all.
– Says one of the former employees of Google.

In addition to literally torturing Bard team members under these circumstances, the chatbot itself can provide false and even dangerous information under these circumstances. The speed at which the moderators are ordered to control the content does not correspond to the high-quality work of the neural network.

According to internal guidelines, team members should check the chatbot’s responses for accuracy, accuracy of details, freshness of information, etc. are requested to check. At the same time, verification should be based solely on the employee’s personal experience or pure “Google”. We remind you that you have 3 minutes for everything.

As a result, the source cites the salaries of people who teach at the Bard – they get $14 an hour, which is extremely low for the United States. At the same time, they are often forced to work again.