Google CEO Sundar Pichai told staff the generative artificial intelligence bot Gemini’s responses to user queries about race, politics, and current events were “completely unacceptable.”
Pichai addressed Google’s team on Tuesday in a memo on the state of the Gemini chatbot and image generator, which have generated controversy over the past week. Google had to shut down the image generator following outrage because it declined to depict white people and inserted minorities into inappropriate situations, such as images of the Founding Fathers, the pope, or vikings. The chatbot has also come under fire for biased responses on several topics relating to race, gender, and politics, such as the Israel-Hamas war and the origin of COVID-19. Republicans have cited the results as evidence of how “woke” Google is.
“I know that some of its responses have offended our users and shown bias — to be clear, that’s completely unacceptable, and we got it wrong,” Pichai said, according to a memo reported by Semafor.
Google’s developers are seeing a “substantial improvement on a wide range of prompts,” Pichai claimed. However, the company will still be creating several updates and reviews to the project before allowing the chatbot to post pictures of people again. These updates include “structural changes, updated product guidelines, improved launch processes, robust evals and red-teaming, and technical recommendations.”
“Our mission to organize the world’s information and make it universally accessible and useful is sacrosanct,” Pichai emphasized. “We’ve always sought to give users helpful, accurate, and unbiased information in our products. That’s why people trust them. This has to be our approach for all our products, including our emerging AI products.”
In response to the controversy over the images generated by Gemini, the company’s AI product lead apologized, and the company paused image generation on Feb. 22.
CLICK HERE TO READ MORE FROM THE WASHINGTON EXAMINER
Critics have claimed it has a left-wing bias or a false understanding of certain topics. These include responses alleging that the results of Hamas’s Oct. 7 attack on Israel are “disputed,” arguing there is no evidence the COVID-19 virus leaked from a laboratory in Wuhan, declining to write an argument in favor of having at least four children, and comparing former Federal Communications Commission Chairman Ajit Pai to Adolf Hitler due to the chairman’s repeal of net neutrality, a policy that would regulate internet providers the same way as telecommunication providers.
Google Knowledge and Information Vice President Prabhakar Raghavan released a blog post on Friday noting the product’s error and stating that the company would tune the AI chatbot for these responses.
