Google just paused its newest AI tool called Gemini, formerly Bard, that creates images of people. This is another big mistake, as Google is striving to get on par with new rivals from OpenAI and Microsoft in using generative AI.
Since OpenAI officially launched in November, 2022, Google has been playing catch up. When Google released its chatbot, originally called Bard now nearly over a year ago. At the time of Bard’s release, Google learned the hard way as it shared inaccurate pictures of a planet outside the Earth’s solar system in a promotional video, causing its shares to slide by over nine percent.
This recent debacle with Gemini had users flagging on social media that Gemini was returning inaccurate historical images. This unfortunately is another example of Google rushing to market without sufficient testing.
In their official communication, they stated this wasn’t what they had intended, and they did not want Gemini to refuse to create images of any particular group. They stated clearly they did not want Gemini to create inaccurate historical — or any other — images. So the disappointing outcome was Google has now turned the image generation of people off and will work to improve it significantly before turning it back on. They learned a tough lesson that this process must include more rigourous and extensive testing.
“Historical contexts have more nuance to them and we will further tune to accommodate that,” said, opens new tab Jack Krawczyk, Senior Director of Product for Gemini at Google, after the social media bantering.
Building trust and confidence in using AI requires careful quality controls. This is an area that many of the technology titans in their haste to bring new innovations to market are risking reputational brand profile, not only is this true for Google with its second misstep with Gemini, but also for the entire AI industry.
Note: The official communication from Google can be found here
