‘Woke to absurdity’ The Google executive responsible for the AI chatbot Gemini says in a tweet that ‘white privilege is real’ and that ‘terrible racism’ is rampant in America. He has been criticized for saying that he is.
The tweets are said to have political content. Written by Jack KlacikA day after the Post reported on Gemini’s strange habit of generating “diverse” images that are historically or factually inaccurate when users request photos of Vikings or American Founding Fathers, Gemini Experience Inc. Gemini Experience, senior director of product at , resurfaced on Thursday’s X.
By Thursday afternoon, Krawczyk had set the X account to private and removed any mention of Google or his role with the company from the account’s profile.
Krawczyk, 40, has set his X-feed to private, but screenshots of his alleged tweets, most of which were created before he was hired by Google in 2020, clearly indicate that he’s super advanced. prejudice has been revealed.
“White privilege sucks,” Krawczyk allegedly wrote in a tweet dated April 13, 2018, according to a screenshot of a post circulating on X. It’s about recognizing gross prejudice at all levels. ”
On January 20, 2021, Mr. Krawczyk commented on President Biden’s inaugural address, “acknowledging systemic racism” and “the American ideal is the dream of the world, but we must work hard to reap the benefits.” He is said to have said that it was “one of the greatest speeches of all time.” that. “
in Another post on October 21, 2020 Apparently after voting against Donald Trump in the presidential election, Kracik allegedly wrote: Bridging the Biden-Harris line felt cathartic. ”
The Polish-born technologist also described America as a place where “racism is the number one value that the people are trying to protect above all else,” and said in other screenshots circulating online that “this “There is clearly severe racism in the country,” he is quoted as saying. Social media.
Attempts to contact Krawczyk directly were not immediately successful.
Google temporarily disabled Gemini’s image generation tool on Thursday after the scuffle, and Krawczyk acknowledged in a statement that it had “missed the mark” by producing revisionist images.
Google declined to comment on the tweet, referring to the company’s previous statement regarding its decision to “pause” Gemini’s image generation tool.
Critics were quick to suggest that Krawczyk’s alleged personal biases may have contributed to Gemini’s tendency to prioritize “diverse” results over accuracy.
“To those responsible for Google’s Gemini AI. And you’ll wonder why they discriminate against white people.” said @LeftismForUThis is an account with a lot of screenshots.
“A woke, race-obsessed moron is the head of product at Gemini,” declared Ian Miles Chong, an influencer who frequently interacts with Elon Musk at X.
Musk personally commented on the tweet, explaining: Google employees as “holes” and “Racist Douche Nozzle”
Musk also shared the Post’s cover story about Gemini’s strange image, writing: “The awakened mind virus is killing Western civilization. Google is doing the same with search results. Facebook and Instagram. And Wikipedia.”
Krawczyk was described as Google AI’s “Teacher” in Men’s Health Profile Since December of last year.
The newspaper could not immediately confirm the authenticity of all the screenshots, but it does say the alleged tweets went viral.
When I searched for the phrase “Jack Krawczyk white privilege,” his current thread appeared as the first result.
When the strange behavior of Gemini’s image generation tool surfaced on Wednesday, Krawczyk was the first Google employee to comment on the issue, saying his team was “working to immediately improve this kind of depiction. ” declared in a statement.
“Gemini’s AI image generation generates a wide variety of people, which is generally a good thing since people all over the world are using it. But we miss the point here,” Krawczyk wrote in the post. told the paper.
Because Google hasn’t published the model that controls the Gemini chatbot’s behavior, it’s difficult to determine exactly what caused the software to fabricate different versions of historical figures and events.
Google’s “training process” with human input is one plausible explanation for this behavior, said Fabio Motoki, a lecturer at the University of East Anglia in the UK. Fabio Motoki is a co-author of a paper last year that discovered a significant left-leaning trend in ChatGPT.
“Reinforcement learning from human feedback (RLHF) recognizes that humans tell a model what is good and what is bad, and actually shape the model’s ‘reward’ function, or more precisely, its loss function. Don’t forget,” Motoki told the Post.
“So depending on what kind of people Google hires or what instructions Google gives them, this could be an issue.”
Krawczyk has been with Google for four years. According to his LinkedIn account, he previously worked on his program for Google Assistant.
Prior to joining Google, he was Vice President of Product Management at WeWork and held positions at VSCO and United Masters.
