Google CEO says 80,000 workers tested Bard A.I., warning ‘Things will go wrong’

Sundar Pichai, CEO of Google and Alphabet, informed staff that the success of the newly established Bard A.I. initiative is dependent on public testing.

“They will surprise us when more people begin to utilize Bard and test its possibilities. Things will go wrong,” Google CEO Sundar Pichai warned in an internal memo to staff obtained by CNBC on Tuesday. Nonetheless, consumer input is vital for enhancing the product and underlying technology.

After months of anticipation, Google debuted Bard as a “experiment” on Tuesday morning, prompting the letter to staff. The tool, which is based on Google’s LaMDA (Language Model for Dialogue Applications), can provide conversational solutions to complex or open-ended inquiries, such as “give me suggestions on how to introduce my daughter to fly fishing.”

Following the announcement, shares of Alphabet increased by nearly 4% at midday.

In many product disclaimers, the business states that Bard may make errors or “provide erroneous or unsuitable replies.”

The most recent internal communication comes as the firm attempts to stay up with the rapidly advancing advances in generative AI technology over the past several months, particularly OpenAI, which is supported by Microsoft, and its ChatGPT technology.

Workers and investors questioned Google after Bard’s original announcement in January, which looked hurried to compete with Microsoft’s recently-integrated ChatGPT into Bing. During a recent all-hands meeting, the most frequently asked questions by workers concerned the function of Bard. At the discussion, officials defended Bard as an experiment and attempted to differentiate the chatbot from the company’s primary search offering.

Pichai’s email on Tuesday also stated that 80,000 Google employees participated to testing Bard in response to Pichai’s all-hands-on-deck call to action last month, which included a request for workers to re-write the chatbot’s incorrect responses.

Pichai stated in his email on Tuesday that the business is attempting to test responsibly and has invited 10,000 trustworthy testers “from all backgrounds and viewpoints.”

Pichai said that staff “should be proud of this effort and the years of technological advancements that have gotten us to this point, including our 2017 Transformer research and foundational models like as PalM and BERT.” Even with all this success, we are still at the beginning stages of a long Al trip, he continued.

“For now, I’m thrilled to see how Bard inspires greater creativity and curiosity in its users,” he said, adding that he is looking forward to discussing “the breadth of our AI accomplishments” at Google’s annual developer conference in May.