Developer conference focuses on features such as image recognition
11 hours ago by: Richard Waters in Mountain View
Google doubled down on artificial intelligence at the company’s developer conference on Wednesday, using its main annual showcase to highlight the technology’s growing importance to its most widely-used services. “In an AI-first world we are rethinking all our products,” said Sundar Pichai, chief executive.
New capabilities added to a series of the company’s services included Google Lens, an image-recognition feature that uses a smartphone’s camera to identify objects in the real world and reveal more information about them.
Google has sought to use AI to make up lost ground in important new technology markets that have been pioneered by rivals, from Amazon’s voice-activated home speaker, Echo, with its virtual assistant Alexa, to Facebook’s messaging services. Since he launched the company on its AI path a year ago, Mr Pichai claimed a number of behind-the-scenes advances had boosted the accuracy of the technology and made new applications possible.
« Google needed to urgently respond to Amazon’s stealth takeover of the home and beyond with Alexa,” said Geoff Blaber, an analyst at CCS Insight. “Broadening availability of Google Assistant and diversifying its feature set is a much needed move as it seeks to use its scale, AI and search assets to close the gap. »
The search company also disclosed that its Android mobile operating system has become the first computing platform to reach 2bn users, a milestone Mr Pichai said it passed this week. Android is one of seven Google products that have passed 1bn users, including Search, Maps and YouTube, giving it wide reach as it tries to find an audience for its new “smart” services.Geoff Blaber, CCS Insight Google first tried using images to power its search engine seven years ago. It abandoned that initiative, called Goggles, after four years when users failed to take up the service. Its return to the idea now follows rapid advances in AI that it says have made ideas like this far more practical. “The fact that computers understand videos and images has profound implications for our core mission,” Mr Pichai said.
In other signs of how deeply AI is reshaping the company, he said Google was reorganising internally to reflect its deepening focus on the technology, bringing together its AI research, infrastructure and application efforts into separate groups. It also disclosed that it had extended its chip design efforts to develop processors capable of “training” advanced AI systems in its own datacentres, putting into more direct competition with chip companies like Nvidia.
Amazon has moved quickly in recent months to cement its early lead in voice-activated computing, and its Alexa service has been embedded in a wide range of hardware. The ecommerce company has also added to the Echo hardware twice in a matter of weeks in a race to turn it into an all-purpose household assistant, first by adding a camera and then unveiling a version with its own screen.
Google responded on Wednesday with an expanded range of partnerships for its rival Home device, and announced a way for the speaker to send information to a user’s existing screens, including smartphones and TVs. Related article Why Britain’s homegrown AI talent leads the world UK expertise is envied by high-tech gurus from Silicon Valley to China
Google’s latest developer conference marked a departure from last year, when it used the event to announce a host of new products and services, including Home, its smart digital assistant, a new mobile virtual reality platform, and new messaging and mobile video conferencing apps. It was accused at the time of announcing “vapourware” — products that only existed on the drawing board. That broke with the company’s normal practice of only showing off new technologies when they are ready for use.
By contrast, this year brought a focus on the core technology that Google hopes will eventually give it an edge in all of these new markets.
In the only new product announcement at the event, Google said it was working on the design for a standalone virtual reality headset that would not require a smartphone, with the first products due for launch by other hardware makers before the end of this year.