"Google Search Is AI Search": Glasses Are Back, Gemini Gets a Brain Upgrade, and 900 Million Users Can't Be Wrong — But DeepMind Workers Vote to Unionize Over Pentagon Deal

20 May 2026 | Mountain View & London — Updated 23:30 GMT

MOUNTAIN VIEW – Google declared Tuesday that "Search is AI search" as it unveiled the most dramatic transformation to its flagship product in nearly three decades — a Gemini-powered overhaul that turns the humble search bar into a personal assistant capable of booking trips, managing spreadsheets, and creating autonomous "information agents."

At the annual Google I/O developer conference, executives also announced that smart glasses are back. More than a decade after wearers of the original Google Glass were dubbed "glassholes" and laughed out of San Francisco, the company is partnering with Warby Parker and Gentle Monster to launch voice-activated, camera-equipped eyewear this fall.

But beneath the surface of product launches and optimism, a different story was unfolding 5,000 miles away. Google DeepMind workers in the UK have voted to unionize — the first "frontier" AI lab to seek union recognition — citing profound ethical concerns over a classified Pentagon deal signed just days earlier.

'Google Search Is AI Search': The Biggest Change in 30 Years

"Google Search is AI search," declared Elizabeth Reid, who oversees the product, calling the changes the biggest Google Search had seen in its nearly 30-year history. The updates rolled out globally Tuesday to desktop and mobile versions.

The search box will now accommodate longer, more specific queries — questions more like those people would ask one another than Search's idiosyncratic syntax. Results will resemble visually augmented versions of AI Overviews, the bulleted summary responses that Google Search currently returns. Users can still choose the original version — the classic list of links — by clicking a tab labeled "Web."

Reid said Google Search queries reached an all-time high last month. Since the debut of Search's "AI Mode" a year ago — in which users chat with a Gemini-powered bot rather than navigating to a list of links — queries to the chatbot-specific interface have doubled every quarter. If users add photos, videos or documents to the Chrome browser's search bar, Search will automatically go into AI mode.

Gemini 3.5: Smarter, Faster, and Now with 'Information Agents'

The overhaul is underpinned by Gemini 3.5, Google's new artificial intelligence model announced the same day. Josh Woodward, who oversees the flagship AI's development, said the goal is "a universal assistant that's personal, proactive, persistent."

Google announced that 900 million people use the Gemini app monthly — still less than ChatGPT's 900 million weekly users, but a significant milestone.

For subscribers to its AI Pro and Ultra plans, Search will include the option to create "information agents" — autonomous AI bots that can perform in-depth research and return summaries or action plans. A "generative user interface" feature will create customized visuals and interactive elements like dashboards based on user requests.

A new agent feature, Gemini Spark, will access information from users' Gmail, Google Calendar and other Google products to perform research, shop, and plan upcoming trips, appointments or recurring tasks. It marks a significant expansion of AI agents beyond tech-savvy, business-focused users to the wider consumer realm.

Glasses Are Back: Warby Parker, Gentle Monster, and a Second Chance

In conjunction with Samsung and eyewear makers Warby Parker and Gentle Monster, Google unveiled "intelligent eyewear" — a line of sunglasses and eyeglasses set to go on sale this fall. The glasses will respond to users' voice commands to Gemini and take photos or videos with embedded cameras.

Fellow tech giant Meta has partnered with EssilorLuxottica to sell similar voice-activated, camera-equipped glasses — the Ray-Ban Meta — since 2021. But Google's history with smart glasses is checkered. In 2014, the company began selling Google Glass for $1,500, inspiring widespread derision. It was shuttered the following year.

In a sign of shifting attitudes, Meta sees its augmented-reality glasses as a major growth area, and Google is working on a second model with an in-lens display, nicknamed Project Aura. Other I/O announcements included AI coding features for Gemini and cybersecurity-focused AI tools.

A Google logo is displayed outside an office building on 12 December 2025 in San Diego, California.

Behind the Curtain: DeepMind Workers Vote to Unionize

While Google executives celebrated product launches on stage, a different battle was being fought in the UK. Workers at Google DeepMind — the company's AI research laboratory — have voted to unionize, seeking recognition from the Communication Workers Union and Unite the Union as joint representatives.

At least 1,000 workers will be represented if the company recognizes the union, according to union officials. It marks the first time a "frontier" AI lab has sought union recognition.

The timing is significant. Just days before the I/O conference, Google reportedly signed a classified AI deal with the US Pentagon, allowing the military to use Google's AI models for "any lawful government purpose" on classified networks. The agreement requires Google to help adjust its AI safety settings and filters at the government's request.

While the contract includes language stating that the AI system is "not intended for, and should not be used for, domestic mass surveillance or autonomous weapons (including target selection) without appropriate human oversight and control," the agreement also says it does not give Google the right to control or veto "lawful" government operational decision-making.

'We Don't Want to Facilitate a Genocide': Workers Speak Out

One DeepMind worker, who requested anonymity because of fear of retaliation, said the union drive was driven by reports of the Pentagon deal, pointing to the US's "capricious Iran war" and the Trump administration's feud with Anthropic as indications the defense department was "not a responsible partner."

Another worker said many at the company had struggled with what they viewed as complicity in Israel's war in Gaza. Google provided the Israeli military with increased access to its AI tools from the early days of the war, the Washington Post reported last year, and in 2021 signed a $1.2bn cloud-computing contract with the Israeli government alongside Amazon.

"Our technology helped the IDF," said the second UK worker. "I want AI to benefit humanity, not to facilitate a genocide."

A Google spokesperson told the Guardian: "Google UK recently received a letter from Unite and the Communications Workers Union requesting recognition for Google DeepMind UK employees. At this stage in the process, there has been no vote to unionize. We have always valued constructive dialogue with employees." A union official reiterated that a vote took place and that there may be a separate ballot for recognition in the future should the company refuse to voluntarily recognize the union.

600+ Employees Sign Open Letter to Sundar Pichai

Last week, amid reports of the pending Pentagon deal, more than 600 Google employees signed an open letter to CEO Sundar Pichai, demanding the company not make its AI systems available for classified use.

"We feel that our proximity to this technology creates a responsibility to highlight and prevent its most unethical and dangerous uses," they wrote. "Therefore, we ask you to refuse to make our AI systems available for classified workloads."

Workers who voted to join the union said they want Google to commit not to develop technology "whose primary purpose is to cause harm or injury to people," establish an independent ethics oversight body, and grant workers the individual right to refuse to contribute to projects on moral grounds. Should the company refuse, they said, they are considering protests and "research strikes" — abstaining from work expected to significantly improve core products like Gemini while continuing less significant updates to avoid detection.

Investors have also raised concerns. A coalition of shareholders who own about $2.2bn of Alphabet's shares wrote a letter last week demanding a meeting and greater transparency about Google Cloud and AI deployments in "high-risk" contexts, including Project Nimbus and services to US immigration authorities.

The Pentagon Deal: 'Any Lawful Government Purpose'

On Friday, the Pentagon confirmed it had reached agreements with seven leading AI companies: Google, SpaceX, OpenAI, Nvidia, Reflection, Microsoft and Amazon Web Services. Anthropic — whose technology is in wide use by the US military but which has sparred with the Pentagon over future contracts — was notably absent.

"These agreements accelerate the transformation toward establishing the United States military as an AI-first fighting force and will strengthen our warfighters' ability to maintain decision superiority across all domains of warfare," defense department officials said in a statement.

The Trump administration has pushed AI companies to make their tools available on classified networks without the standard restrictions they apply to users. Google's contract reportedly includes non-binding language about not using AI for mass surveillance or autonomous weapons without human oversight. But the agreement also states Google has no right to control or veto "lawful" government operational decision-making.

The Pentagon has said it has no interest in using AI to conduct mass surveillance of Americans or to develop lethal weapons that operate without human involvement, but wants "any lawful use" of AI to be allowed.

This is not Google's first military controversy. In 2018, thousands of employees protested Project Maven, in which Google agreed to build AI products for Pentagon drone footage analysis. The company did not renew the contract in 2019 and published AI principles that included a pledge — since dropped — not to design AI for weapons. Palantir took over Project Maven, which continues today.

Last year, Alphabet lifted its ban on AI for weapons and surveillance, removing language promising not to pursue "technologies that cause or are likely to cause overall harm." AI lead Demis Hassabis said in a blogpost that AI had become important for protecting "national security." Some employees expressed concerns on internal message boards. One asked: "Are we the baddies?" according to Business Insider.

Stay updated with the latest tech trends on our Tech Trends Page.