SOMETIMES it is perceived as a figment of the far future. But artificial intelligence (AI) is today’s great obsession in Silicon Valley. Last year technology companies spent $8.5 billion on deals and investments in artificial intelligence, four times more than in 2010. Nearly all of the world’s technology giants, including Google, Microsoft, Facebook, Amazon and Baidu, are competing fiercely to hire the best AI experts, snap up start-ups and pour money into research. What accounts for the tech elite’s sudden AI-phoria?

The technology has not always been so popular. The field was largely ignored and underfunded during the “AI winter” of the 1980s and 1990s. At that time AI research conducted at universities proved to be disappointingly slow and irrelevant to companies’ bottom lines. Now, however, the chill is gone. Progress in AI is accelerating. Recently Google generated lots of headlines when DeepMind, a start-up it acquired in 2014, helped train a computer to repeatedly beat the world champion at Go, a board game. This has sparked both fear and hope for the future of AI: hope for fat profits and improving people’s lives through technology; fear about how society will cope with the dislocation AI could bring.  

AI is already starting to generate big financial gains for companies, which helps explain firms’ growing investment in developing AI capabilities. Machine-learning, in which computers become smarter by processing large data-sets, currently has many profitable consumer-facing applications, including image recognition in photographs, spam filtering and those that help to better target advertisements to web surfers. Many of tech firms’ most ambitious projects, including building self-driving cars and designing virtual personal assistants that can understand and execute complex tasks, also rely on artificial intelligence, especially machine-learning and robotics. This has prompted tech firms to try to hire up as much of the top talent as they can from universities, where the best AI experts research and teach. Some worry about the potential of a brain drain from academia into the private sector.

The biggest concern, however, is that one firm corners the majority of the talent in artificial intelligence, creating an intellectual monopoly of sorts. Google looks best positioned to do this: between its Google Brain project and its acquisition of DeepMind, it has some of the brightest human brains working on AI. Because superior AI systems are able to learn and improve more quickly, the firms that develop an early edge in artificial intelligence may reap the greatest rewards and erect barriers to entry that smaller firms will find hard to overcome. In December Elon Musk and several other tech leaders pledged $1 billion to help fund OpenAI, a research lab that will make public all of its findings, to ensure there is an entity that is working on developing AI on behalf of the public good and not just its own profits. Today AI is the domain of tech geeks, but its future matters to everyone.

in 2007 Deep Learning was born now in Feb 2016 it’s part of Search

In 2007, in the midst of this AI work, a Google Tech Talk in Mountain View about Deep Learning, which galvanized the geeks in attendance, and won a huge following on YouTube. It helped spread the news that neural nets were finally going to be a powerful tool. And the rush was on to hire people who understood this new technique. Hinton’s students went to IBM, Microsoft, and, of course, Google. That represented three of the four major companies working in the field (the other one, Nuance, includes Apple among its suppliers). All were free to use the work from Hinton’s lab in the systems each would help refine in his respective company. “We basically gave it away because we were very concerned to prove we had the goods,” says Hinton. “What was interesting was that MSR [Microsoft Research] and IBM got it before Google but Google turned it into a product faster than anyone else.”

EARLIER this year BMW advertised on WeChat, a popular messaging app in China with around 550m monthly users. But its ads were shown only to those whose profiles suggested they were potential buyers of expensive cars. Others were shown ads for more affordable stuff, such as smartphones. The campaign bruised a few egos. Some of those not shown the BMW ad complained, referring to themselves as diao, or (putting it politely) losers.

The carmaker’s experience shows the complexities of advertising today, when it is so easy for dissatisfied customers to make their voices heard. But it was also an example of how marketing chiefs are struggling to find the right way to reach consumers on new digital platforms, where they are spending ever more of their time.

Not long ago social-media marketing was something that brand managers might ask their summer interns to deal with. Today it has become a pillar of the advertising industry. Social networks like Facebook, Twitter and LinkedIn have cultivated vast audiences: 2 billion people worldwide use them, says eMarketer, a research firm. Online advertising of all sorts continues to grow, and within that category, spending on social-media ads has gone from virtually nothing a few years ago to perhaps $20 billion this year (see charts).

Advertisers like social-media platforms because they gather all sorts of data on each user’s age, consumption patterns, interests and so on. This means ads can be aimed at them with an accuracy that is unthinkable with analogue media. For example, Chevrolet, an American car brand, has sent ads to the Facebook pages and Twitter feeds of people who had expressed an interest in, or signed up to test-drive, a competitor’s vehicle.

Such fine-tuned targeting means that the distinction between advertising and e-commerce is becoming blurred. Facebook, Twitter, Instagram and other platforms are selling ads containing “buy now” buttons, which let users complete a sale on the spot. It is too early to tell how many consumers want such a convenience, but the social platforms foresee a future in which they get paid by advertisers to provide instant-shopping services that make the platforms more useful to their members, and get them to spend more time on them.

To wring the most out of the ability to target consumers precisely on social media, ad agencies are making big changes to their campaigns. Instead of creating a single, broad-brush message that will run across television, radio, print and outdoor, they are producing many variations on a theme, matching each to the subset of consumers they judge most likely to respond to it. Last month Lowe’s, an American home-improvement retailer, ran a campaign on Facebook in which users were sent one of several dozen versions of its ad, depending on which part of their homes they had mentioned on social media.

The iterative nature of digital marketing has meant lots of work for ad agencies and public-relations firms. However, the brands that hire them must weigh the “production-cost trade-offs” that come with personalisation on social media, says Pete Blackshaw of Nestlé, a food manufacturer. “You can target too much.”

Marketing chiefs also need to think through efforts to give their brands “online personalities”. Twitter has been the main place where brands try to sound authentic and clever. When Apple announced its gold iPhone in 2013, Denny’s, a restaurant chain, sent out a gently mocking tweet showing a photo of its pancakes, which are “always available in golden”. It seemed to go down well with consumers.

But in attempting to ride social-media trends, companies can easily fall flat on their faces. DiGiorno, a frozen-pizza maker, noticed that a number of people were using the hashtag “#WhyIStayed” on Twitter and sent out a jocular tweet that they had stayed for the pizza. It turned out that the comment thread was about domestic violence and why women remained in abusive relationships. DiGiorno took to social media again—this time to apologise.

In spite of such pitfalls, social platforms are likely to receive an ever larger part of marketers’ budgets. But the digital-media business is still young and volatile, and it is hard to predict which social networks are destined to become the new-media equivalents of America’s big four broadcast-TV networks. For a time, Twitter seemed to be the place to be for advertisers; more recently it has been dogged by management turnover and slowing user growth. Now Facebook is the favourite among marketing folk: it claims nearly five times as many users and nine times as much revenue as Twitter. Facebook has bought several nascent social-media services that might have grown to become challengers, such as Instagram, a photo-sharing app, and WhatsApp, a messaging app. It has been rolling out ads cautiously on Instagram, to see how users react, but has yet to start doing so on WhatsApp.

Some old media have yet to feel much pain from the loss of ad revenue to digital rivals. TV advertising has until now kept growing. But as time goes on, and as TV audiences both decline and shift to services that do not have ads, such as Netflix, the competition will be more keenly felt. However, social networks, and TV advertisers interested in switching to them, have yet to work out what is the optimal format for video ads. In 2012 Twitter acquired Vine, which lets people post six-second videos; several months ago Periscope, an app for live video also owned by Twitter, was all the rage. Advertisers have experimented with both services, but as yet neither has taken off as a marketing medium.

It still makes sense for marketers to try these new services out, because there is something of a first-mover advantage in digital advertising. Brands that are early to use new platforms benefit disproportionately, explains Linda Boff, a marketing chief at General Electric, because their users have not yet become saturated with marketing messages.

The latest social platforms to get the attention of marketing types are a bunch of messaging apps, such as Snapchat (see article), WeChat and Kik, where young users send messages, photos and videos directly to friends. Brands are also starting to do more with Pinterest, where users can “pin up” images of things that appeal to them. It seems a fair assumption that users may want to buy the things they are pinning up, although the platform, which has 70m users, may never achieve the scale of Twitter (300m), let alone Facebook (1.5 billion).

Even if marketers master social media without coming across as clumsy, grating or intrusive, there will surely be a limit to how much advertising will shift to the platforms. Television ads are still great for reaching big audiences with simple messages. Print ads can lend brands an air of credibility (we would say that, wouldn’t we?). Like fund managers, advertisers will always want a balanced portfolio.