Before launching Bard, Google employees warned that the AI ​​chatbot was not yet ready for the general public. However, ethical considerations were ignored in favor of speed.

Google employees who could work internally with the AI ​​chatbot Bard asked their company not to launch the tool. Bloomberg found this out after interviewing (former) employees, evidenced by screenshots of interviews. Internal testing found Bard to be “rather useless than useless.” Another tester called the system “a pathological liar.”

Potentially dangerous answers

After all, Bard frequently gave wrong answers, which could even lead to damage on a regular basis. For example, when employees asked for diving tips, the answers were a guarantee of trouble and maybe even injury.

Google has internal mechanisms whereby a product is given a score in various categories that indicates its market maturity. Bard did not score highly enough in several categories. Both tests with employees and internal guidelines indicated that a market launch was not yet in sight.

competitive pressure

However, under pressure from OpenAI and ChatGPT, Google decided to ignore these concerns. The solution: label the product “experimental” in the hope that the public will forgive shortcomings. Staffers Bloomberg spoke to were dismayed at how “the trusted search giant was providing inferior information to keep up with the competition.”

The result was painful. When Bard was launched, a lot went wrong and Bard gave incorrect information, among other things, just as the staff had predicted. In response to the presentation, the search giant’s stock plummeted.

Public exams

Not everyone at Google sees speed as a problem. On the contrary, some believe that public testing is a crucial part of generative AI development. The fact is that OpenAI has launched a more mature tool with the launch of ChatGPT, although ChatGPT was and is not bug-free either.

The testimonials mainly show how Google has picked up speed from OpenAI and Microsoft and is afraid of falling even further behind. While a company spokesperson tells Google that ethical considerations remain Google’s top priority, Bloomberg’s research shows that other issues in the AI ​​competition are taking precedence. This in turn shows that it is unrealistic to expect adequate self-regulation from companies in the face of fierce competition.