Kathy Gibson reports – The artificial intelligence (AI) arms race is on: hot on the heels of Microsoft’s addition of OpenAI to its search products, Google has showcased how it is already using generative AI, along with new offerings and its vision for the future.
“Although we are 25 years into search, I dare say we have just begun,” says Prabhakar Raghavan, senior vice-president at Google. “We are reinventing what it means to search, and the best is yet to come.”
Raghavan points out that the key to understanding information that empowers people is language. And, until recently, language was accessible only by the people who understood it.
Today, more than 1-billion people use Google Translate to understand conversations in 133 languages, and the company recently added 33 new languages to Translate’s offline mode.
“And soon we will offer a richer and more intuitive way to translate language which will consider the context needed for the right turn of phrase,” Raghavan says.
Google’s Zero-Shot Machine Translation tool, which can translate from multiple languages, has now added 24 new languages to further bridge the language divide.
“But, although language is at the heart of how we communicate, we also communicate visually,” Raghavan says. “Back in 2017, we redefined search with Lens and have since brought it directly on to the search bar. We recently reached a major milestone of more than 1-billion people using Lens in a month – so we can say the age of visual search is here.”
New developments in Lens allow users to translate text overlaid on pictures, seamlessly replacing the original text – and to translate text in pictures within their context. In the coming months, Lens will allow users to visually search across multiple apps on their mobile phones.
“With multisearch you can search with text and pictures together,” explains Liz Reid, vice-president: search at Google. “This ability lets you mix modalities and opens up a whole world of possibilities.”
The ability to use multisearch to locate places near the user will be rolled out across the globe in the coming months.
Google has had a two-year start on adding generative AI into its services, with next-generation language and conversation capabilities powered by its Language Model for Dialogue Applications (LaMDA).
Its conversational AI service, Bard, is now being opened up to an ecosystem of trusted developers before becoming more widely available to the public in the coming weeks. “We will continue to ensure it meets our high bar of quality and trust,” Raghavan says.
While Google has a long history of offering up the answers to factual queries, Raghavan points out that there are many queries for which there is no right answer, or Nora queries, where users may want to explore a diverse range of opinions. “So we are bringing generative AI directly to your search results, which will help to organise complex responses and opinions, from where the user can quickly explore new angles,” he says.
But the potential for generative AI goes beyond language and text, Raghavan adds. Not only can content creators automate 360-degree animations from just a handful of images, users will be able to search using text, voice, images or a combination of these models.
“As we continue to bring generative AI technologies to our products, the only limit to search will be your imagination,” he says.
Google is opening up its Generative API (application programming interface) to developers and partners and, over time, will create a suite of tools to allow others to create applications using AI.
“When it comes to AI, it is critical that we bring these experienes to the world responsibly,” Raghavan stresses. “We have been focusing on responsible AI since the beginning and working with partners and the development community to ensure that we bring AI to people in a way that is bold and responsible.”
An example of a more immersive experience is the new developments in Google Maps.
“We are evolving the two-dimensional map into a multi-dimension view of the real world,” says Chris Phillips, vice-president and GM: geo at Google. “This immersive view is a more natural way to explore and is truly intuitive.
“It allows the user to truly experience a place before they step inside – and offers a completely new way to interact with the map.”
Search with LiveView has been introduced in some cites and will be rolled out to additional locations over the coming months.
Painting AI with augmented reality is a key to Indoor LiveView, which is in the process of adding about 1 000 new venues.