Sergey Brin says Google ‘definitely messed up’ with Gemini launch

Google co-founder Sergey Brin, in a rare public appearance over the weekend, told a group of artificial intelligence enthusiasts that he came out of retirement “because the trajectory of AI is so exciting.”

Brin, 50, spoke to entrepreneurs on Saturday at the “AGI House” in Hillsborough, California, just south of San Francisco, where developers and founders were testing Google’s Gemini model. AGI stands for artificial general intelligence and refers to a form of AI that can complete tasks to the same level, or a step above, humans.

In taking questions from the crowd, Brin discussed AI’s impact on search and how Google can maintain its leadership position in its core market as AI continues to grow. He also commented on the flawed launch last month of Google’s image generator, which the company pulled after users discovered historical inaccuracies and questionable responses.

“We definitely messed up on the image generation,” Brin said on Saturday. “I think it was mostly due to just not thorough testing. It definitely, for good reasons, upset a lot of people.”

Google said last week that it plans to relaunch the image generation feature soon.

Brin co-founded Google with Larry Page in 1998, but stepped down as president of Alphabet in 2019. He remains a board member and a principal shareholder, with a stake in the company worth about $100 billion. He’s returned to work at the company as part of an effort to help ramp up Google’s position in the hypercompetitive AI market.

In some cases on Saturday, Brin said he was giving “personal” answers, as opposed to representing the company.

“Seeing what these models can do year after year is astonishing,” he said at the event, a recording of which was viewed by CNBC.

Regarding the recent challenges with Gemini that led to flawed image results, Brin said the company isn’t quite sure why responses have a leftward tilt, in the political sense.

“We haven’t fully understood why it leans left in many cases” but “that’s not our intention,” he said. The company has recently made accuracy improvements by as much as 80% on certain internal tests, Brin added.

Brin’s comments represent the first time a company executive has spoken on the Gemini matter in a live setting. The company previously sent prepared statements from Prabhakar Raghavan, Google’s head of search, and CEO Sundar Pichai in response to the controversial rollout.

Here’s what Raghavan said in a blog post on Feb. 23:

“So what went wrong? In short, two things. First, our tuning to ensure that Gemini showed a range of people failed to account for cases that should clearly not show a range. And second, over time, the model became way more cautious than we intended and refused to answer certain prompts entirely — wrongly interpreting some very anodyne prompts as sensitive. These two things led the model to overcompensate in some cases, and be over-conservative in others, leading to images that were embarrassing and wrong.”

Google declined to comment for this story. Brin didn’t immediately respond to a request for comment. 

Brin said Google is far from alone in its struggles to produce accurate results with AI. He cited OpenAI’s ChatGPT and Elon Musk’s Grok services as AI tools that, “say some pretty weird things that are out there that definitely feel far left, for example.”

Hallucinations, or false responses to a user’s prompt, are still “a big problem right now,” he said. “No question about it.”

“We have made them hallucinate less and less over time, but I’d definitely be excited to see a breakthrough that’s near-zero,” Brin said. “But you can’t just like — count on breakthroughs so I think we’re just going to keep doing the incremental things we do to bring it down, down, down over time.”

When asked by an attendee if he wants to build AGI, Brin answered in the affirmative, citing the ability for AI to help with “reasoning.”

Brin was also asked how online advertising will be disrupted considering ad revenue is core to Google’s business. The company has reported slowing ad growth in the last few years.

“I of all people am not too terribly concerned about business model shifts,” Brin said. “I think it’s wonderful that we’ve been now for 25 years, or whatever, able to give just world class information search for free to everyone and that’s supported by advertising, which in my mind is great for the world.”

He did acknowledge that the business is likely to change.

“I expect business models are going to evolve over time,” he said. “And maybe it will still be advertising because advertising could work better, the AI is able to better tailor it.”

Brin is confident in Google’s position.

“I personally feel as long as there’s huge value being generated, we’ll figure out the business models,” he said.

Beyond AI, Brin was asked about Google’s difficulties in hardware given recent advancements in virtual reality. Google was notoriously early to the augmented reality market with the now-defunct Google Glass.

“I feel like I made some bad decisions,” he said, referring to Google Glass. If he were doing it differently, Brin said, he would have the treated Google Glass as a prototype instead of a product. “But, I’m still a fan of the lightweight” form, he said.

In regards to the Apple Vision Pro and Meta’s Quest headsets, Brin said, “They’re very impressive.”

When asked about how he sees Gemini impacting spatial computing or products like Google Maps or Street view, Brin responded with as much curiosity as anything.

“To be honest, I haven’t thought about it, but now that you say it, yeah there’s no reason we couldn’t put in more 3D data,” Brin said, to laughs from the crowd. “Maybe somebody’s doing it at Gemini — I don’t know.”

WATCH: Google vs. Google