Even the employees don’t think so. While Google, for its part, introduced its chatbot, called Bard, on March 21, Bloomberg reveals that the employees have asked that it not be released.
The survey sheds light on how Google was able to sidestep ethical issues to stay in the race against rivals like Microsoft and OpenAI.
For its investigation, the US agency relies on testimonials from 18 current or former Google employees and screenshots of internal messages. To back up his point, the US media cites the example of an employee: when he asked Bard how to land a plane, Bard regularly gave him suggestions that would lead to an accident. Another employee asked about diving, to which Bard replied “that it would likely result in serious injury or death.”
“Do not throw it away”
Google voluntarily doubled its AI team in 2021 and devoted more resources to assessing the risks of these AI technologies. But interest in innovation forced Google to speed up. In early 2021, the company even fired researchers Timnit Gebru and Margaret Mitchell, who exposed flaws in AI language systems, particularly those used by Bard.
Google also reportedly avoided a risk assessment requested by an internal security team. According to the latter, Bard was not ready for general use. Google still opened experimental access in March.
“We continue to invest in the teams working to apply our AI principles to our technology,” Google spokesman Brian Gabriel said, adding that AI ethics remains “a top priority for the company.”
Source: BFM TV
