Google has unleashed another wave of artificial intelligence designed to tackle more of the work and thinking done by humans as it tries to stay on the technology’s cutting edge.

The next generation of Google’s AI is being packaged under the Gemini umbrella, which was unveiled a year ago. Google is framing its release of Gemini 2.0 as a springboard for AI agents built to interpret images shown through a smartphone, perform a variety of tedious chores, remember the conversations consumers have with people, help video game players plot strategy and even tackle the task of doing online searches.

In a blog post, Google CEO Sundar Pichai predicted the technology contained in Gemini 2.0 will “understand more about the world around you, think multiple steps ahead and take action on your behalf, with your supervision.”

A lot of Google’s latest AI technology will initially be confined to test groups and subscribers who pay $US20 per month for Gemini Advanced, but some features will be made available through its search engine and mobile apps.

Google is planning wider releases next year that will include the technology popping up in its smorgasbord of free products, including its Chrome browser, digital maps and YouTube.

Besides trying to outshine OpenAI and other ambitious startups, Google is also trying to stay a step ahead of Apple as that trendsetting company begins to blend AI into its latest iPhones and other devices.

After releasing a software update enabling the first bundle of the iPhone’s Apple Intelligence features that spruced up the device’s Siri assistant, another batch of the AI technology came out with a free software update that was also released on Wednesday.

Google is pushing forward with its latest AI advances even as the US Justice Department is trying to break up the company to prevent further abusive practices by its dominant search engine, which was declared an illegal monopoly by a federal judge earlier this year as part of a landmark antitrust case.

Among other things, Gemini 2.0 is supposed to improve the AI overviews that Google began highlighting in its search results over its traditional listing of the most pertinent links to websites earlier this year in response to AI-powered “answer engines” such as Perplexity.

After the AI overviews initially produced some goofy suggestions, including putting glue on pizza, Google refined the technology to minimise such missteps.

Now, company executives are promising things are going to get even better with Gemini 2.0, which Pichai said will be able to engage in more human-like reasoning while solving more advanced math problems and even churn out some computer code. The improvements to AI Overviews will initially only appear to a test audience before a wider release next year.

As part of Gemini 2.0, Google is also going to begin testing an extension to Chrome called Project Mariner, which can be turned on to do online searches and sift through the results so people don’t won’t have to bother.

If the US Department of Justice gets its way, Google will be forced to sell or spin off Chrome as part of its punishment for deploying its search engine in ways that stifled competition and potential innovation. Google has ridiculed the proposal as “overly broad” and vowed to resist any attempt to break up the company during federal court hearings scheduled to begin in Washington next  year.

 

Michael Liedtke
(Australian Associated Press)