Information technology Google wants to launch its answer to Meta's AI glasses on the market

SDA

19.5.2026 - 23:19

A year ago, Google presented prototypes of glasses with a small display in one of the lenses. However, a simpler model with a camera and loudspeakers will be launched on the market first. (archive image)
A year ago, Google presented prototypes of glasses with a small display in one of the lenses. However, a simpler model with a camera and loudspeakers will be launched on the market first. (archive image)
Keystone

In the race of AI developers, Google is relying on its own strengths such as the leading web search engine. The internet giant wants to attract users to its services with artificial intelligence even more via search.

Keystone-SDA

Many of the more powerful new capabilities of Google's Gemini AI are behind a paywall and are initially mainly available in the USA.

Google is continuing to develop its search mask, which has hardly changed for decades, for the era of artificial intelligence (AI). Among other things, the search field will now expand as required to allow longer queries. In addition to text, it can also be filled with images, files or videos. In addition, AI will go beyond the previous auto-complete suggestions when entering questions and do much more.

In recent years, there has been much speculation as to whether new rivals that use artificial intelligence to provide answers to queries rather than links could make life difficult for Google's leading search engine. However, the internet giant itself has integrated such functions as summaries created by artificial intelligence - and emphasizes that the search function is used more as a result. At the same time, some website operators are complaining that fewer users are now coming to them via Google.

Google search is now an AI search through and through, announced the responsible manager Liz Reid at the Google I/O developer conference. In addition to summaries, the search engine will also be able to generate explanatory graphics for complex topics or create a fitness plan. Ideally, it will find the right individual format for every query, said Google manager Robby Stein.

"Information agents" to trawl the web

Overall, Google is now focusing heavily on artificial intelligence. The announcements at I/O also include new so-called AI agents: Programs that can independently perform a sequence of tasks for users. Here, Google is competing with AI companies such as ChatGPT inventor OpenAI and Anthropic.

Google is introducing "information agents" in search, for example. They are designed to continuously search the web for predefined information and provide users with tips. For example, if you are looking for an apartment with certain specifications, a Google agent could let you know when something suitable comes up. Gemini will become a "universal assistant", said Google manager Josh Woodward.

How Google wants to transform online shopping

Google also has ambitious plans for online retail. Since the invention of AI agents, there has been a vision that they could independently go online to buy things for users. Google wants to play a decisive role in shaping the infrastructure for this.

For example, the company presented an "intelligent" digital shopping basket that will work across various retailers and services. Among other things, it will automatically search for lower prices and special offers for selected items - or inform you when something that is sold out is available again. And when you put together components for a computer, for example, this shopping basket should issue a warning if they are not compatible with each other.

A technical protocol developed at Google is intended to enable AI agents to pay for automated purchases on behalf of users. This involves specifying which products from which brands are to be purchased - and how much they may cost.

Better recognition of AI content

Google is also making it easier to recognize images created or edited with the help of artificial intelligence - initially in the app for the in-house AI software Gemini, later also in search and in the Chrome web browser. For example, if there is any doubt as to whether a photo or video is authentic, it will be possible to ask whether AI was involved. Google wants to provide other services with an interface for an AI recognition function. Companies such as Nvidia and OpenAI want to use Google's invisible watermark called SynthID for images and other AI-generated content in the future.

Paywalls

Some of the new functions appear behind paywalls. These include the powerful AI agent Spark, which will be able to organize the lives of its users with extensive access to personal data. To be able to use Spark, you need the Ultra AI subscription for at least around 100 euros per month. This pricing is a current trend among all major AI providers. Artificial intelligence requires huge capacities in data centers - and companies cannot keep up with demand despite the expensive expansion.

This also applies to Google, said CEO Sundar Pichai at I/O. There are bottlenecks across the board. The internet giant is countering this with huge investments. Google's parent company Alphabet recently increased the range of expected capital expenditure this year by 5 billion to between 180 and 190 billion dollars, with the prospect of "significantly" more in 2027. Pichai pointed out that, in view of rising prices for components, it is now possible to build less computing capacity for the same amount of money than in the past.