Google CEO angered by totally unacceptable Gemini mistakes

Google CEO angered by “totally unacceptable” Gemini mistakes

Google CEO Sundar Pichai denounced “completely unacceptable” flaws in its artificial intelligence application Gemini in a letter to his teams, saying the creation of historically inaccurate images led to it halting the creation of images of people on Thursday.

• Also read: Chat GPT competitor: Google shares its Gemini in Canada

The controversy arose weeks after Google announced the name change of Bard, its generative AI, to Gemini, a competitor to Microsoft-backed OpenAI ChatGPT.

On social networks, users pointed out certain inaccurate creations, particularly regarding gender and race, that appeared to underrepresent white people. For example, a query about a German soldier from 1943 resulted in images of Asian or black-skinned soldiers.

“I would like to discuss the latest issues related to the generation of problematic (content) in the Gemini application,” Mr Pichai wrote on Tuesday in a letter published by the news site Semafor, which was confirmed by a Google spokesman to AFP.

“I know some of his answers were biased and shocked our users. And to be clear, this is completely unacceptable and we got it wrong.”

Sundar Pichai said Google teams were working “around the clock” to resolve these issues, but did not say when people image generation would be available again.

“No AI is perfect, especially in the nascent phase of this industry, but we recognize that the bar is high and we will continue to work for as long as it takes,” he wrote.

Since the end of 2022 and the success of ChatGPT, generative AI capable of producing text, sounds, images or videos in everyday language upon simple request has generated great excitement and the tech giants are racing to provide tools for Companies and individuals.

But AI models have long been criticized for perpetuating racial and gender biases in their results.

Google said last week that the problematic twin generations stemmed from its efforts to eliminate these biases.

Since the appearance of ChatGPT, which introduced the general public to the potential of AI, many experts and governments have warned of certain risks associated with it, such as the spread of fake, larger-than-life photos, and warned of the danger of opinion manipulation.