Google is known for seemingly wild investments like stratospheric Internet balloons and face computers. This week it began a project that is less flashy but bolder: rethinking the conventional search engine that has become embedded into daily life and provided the revenue that made Google into a $546 billion behemoth.
On Wednesday the company launched a virtual helper similar to Apple’s Siri. Called Google Assistant, the awkwardly named aide exists only as a “preview” inside Google’s new messaging app Allo, and early impressions show that it still needs work. But Google is committed to rolling out Assistant much more widely, in a bid to make us as dependent on it as we are on the search box.
“We are evolving search,” said Google CEO Sundar Pichai when he first described the new feature this summer. “We think of this as building each user their own individual Google.”
For now, Assistant has two roles within Allo. It offers to help out if you’re chatting with a friend about certain topics, such as where to get coffee or what comedy movie to see. Assistant can drop details about the top-reviewed businesses or relevant movies into the message thread.
You can also chat privately to Google Assistant, using text or voice. It can handle queries and commands similar to the ones Siri addresses—for example, if you want to know about tomorrow’s weather or set a timer. But Assistant appears to do better than Siri on factual and map-related queries, thanks to Google’s search technology. Allo can also understand photos. Send it a photo of a dog and it will identify the breed and offer other information, such as its usual life span. Snap the cover of a book and it will present an info box, including the average rating on GoodReads.
The next step in Google’s campaign to reinvent search will come later this year with the launch of Google Home, an Internet-connected speaker closely modeled on Amazon’s Echo. You will be able to call out to the device from across the room to talk with Google Assistant, just as you can to Echo to interact with Amazon’s virtual assistant, Alexa.
After that, Assistant will probably be added to Google’s mobile search apps and tightly integrated into the Android operating system running on more than a billion phones and tablets around the world.
Google has also said that it will let other companies integrate with Assistant, making it possible to do things like call a ride with Uber or book a restaurant. A job ad indicates that Google also wants to strike deals with other companies to help distribute Assistant. That could mean letting manufacturers build Assistant into their own products, such as home speakers or other gadgets, a strategy that Amazon is pursuing with Alexa. Google is also hiring writers to work on the content that will give Assistant its personality.
If all that works, Google Assistant could undermine the company’s business model. If you search for a personal injury lawyer on Google, several ads appear at the top of your results. Asking Google Assistant to “find me a personal injury lawyer” just called up a list of local numbers and a map.
Inserting ads into Assistant’s replies in Allo would ruin the conversational feel. Including ads in responses spoken out loud via Google Home would be even more difficult. When I asked Google’s head of search about this earlier in the summer, he said the company was planning to figure out how to monetize searches made via Assistant after it started being widely used, not before.
A more pressing priority is for Google to prove that Assistant can live up to the hype—something predecessors such as Siri and Alexa have failed to do. Like them, Assistant frequently disappoints by not grasping what you’ve asked. After Google gave the New York Times five days to try out Allo before its release, its writer dubbed Google Assistant an “annoying intern.”
People may learn to work around Assistant’s limitations, as they have learned to translate their desires into “search-ese” to get the most out of Google’s search box. But Google has a big technical challenge ahead of it. We are still a long way from software that can use language anything like the way a human does (see “AI’s Language Problem”).
Google is likely to fight hard to make it work. The depth and ambition of its investment in machine learning is unparalleled, and language is one of the main focuses of that research. Positioning Assistant as the next evolution of search, one of the most successful products of all time, gives the company much stronger motivation than its competitors have. Amazon’s Alexa and Apple’s Siri are side projects. Search is Google’s economic and cultural lifeblood.
This new data poisoning tool lets artists fight back against generative AI
The tool, called Nightshade, messes up training data in ways that could cause serious damage to image-generating AI models.
Rogue superintelligence and merging with machines: Inside the mind of OpenAI’s chief scientist
An exclusive conversation with Ilya Sutskever on his fears for the future of AI and why they’ve made him change the focus of his life’s work.
Data analytics reveal real business value
Sophisticated analytics tools mine insights from data, optimizing operational processes across the enterprise.
Driving companywide efficiencies with AI
Advanced AI and ML capabilities revolutionize how administrative and operations tasks are done.
Get the latest updates from
MIT Technology Review
Discover special offers, top stories, upcoming events, and more.