The sunny and the dark side of AI

MENTION ARTIFICIAL INTELLIGENCE, and most people will think of robots. But a more fitting image may be that of Janus, the Roman god of beginnings, transitions and endings, who has two faces looking in opposite directions. On one side are the positive changes that AI will bring, enabling people to achieve more, far more quickly, by using technology to enhance their existing skills. Recruiters will be able to pinpoint the best candidates more easily, and customer-service staff will be able to handle queries faster. Jobs that never existed before could be created. And getting machines to do routine work can make professional lives more fulfilling and stimulating.

Consumers, too, will benefit from AI-enhanced services such as personalised recommendations and faster and more efficient delivery, as well as from radical changes in industries like health care and transport that could lead to new drug discoveries and treatments and safer ways to move around.

Look the other way, though, and there are plenty of potential pitfalls. Technological change always causes disruption, but AI is likely to have a bigger impact than anything since the advent of computers, and its consequences could be far more disruptive. Being both powerful and relatively cheap, it will spread faster than computers did and touch every industry.

Sunny with a chance of thunderstorms

In the years ahead, AI will raise three big questions for bosses and governments. One is the effect on jobs. Although chief executives publicly extol the broad benefits AI will bring, their main interest lies in cutting costs. One European bank asked Infosys to find a way of reducing the staff in its operations department from 50,000 to 500. This special report has shown that AI-enhanced tools can help pare staff in departments such as customer service and human resources. The McKinsey Global Institute reckons that by 2030 up to 375m people, or 14% of the global workforce, could have their jobs automated away. Bosses will need to decide whether they are prepared to offer and pay for retraining, and whether they will give time off for it. Many companies say they are all for workers developing new skills, but not at the employer’s expense.

A second important question is how to protect privacy as AI spreads. The internet has already made it possible to track people’s digital behaviour in minute detail. AI will offer even better tools for businesses to monitor consumers and employees, both online and in the physical world. Consumers are sometimes happy to go along with this if it results in personalised service or tailored promotions. But AI is bound to bring privacy violations that are seen as outrageous. For example, facial-recognition technology has become so advanced that it may be able to detect someone’s sexual orientation. In the wrong hands, such technology could militate against fair and equal treatment. Countries with a record of surveillance and human-rights abuses, such as China, are already using AI to monitor political activity and suppress dissent. Law-enforcement officials around the world will use AI to spot criminals, but may also snoop on ordinary citizens. New rules will be needed to ensure consensus on what degree of monitoring is reasonable.

Janus, the Roman god, contained both beginnings and endings within him. That duality characterises AI, too

The third question is about the effect of AI on competition in business. Today many firms are competing to provide AI-enhanced tools to companies. But a technology company that achieves a major breakthrough in artificial intelligence could race ahead of rivals, put others out of business and lessen competition. This is unlikely to happen in the near future, but if it did it would be of great concern.

More likely, in the years ahead AI might contribute to the rise of monopolies in industries outside the tech sector where there used to be dynamic markets, eventually stifling innovation and consumer choice. Big firms that adopt AI early on will get ever bigger, attracting more customers, saving costs and offering lower prices. Such firms may also reinvest any extra profits from this source, ensuring that they stay ahead of rivals. Smaller companies could find themselves left behind.

Retailing is an illustration of how AI can help large firms win market share. Amazon, which uses AI extensively, controls around 40% of online commerce in America, helping it build moats that make it harder for rivals to compete. But AI will increase concentration in other industries, too. If, say, an oil company can use AI to pump 3% more efficiently, it can set prices 3% lower than those of a rival. That could force the competitor to shut down, says Heath Terry of Goldman Sachs. He thinks that AI has “the power to reshuffle the competitive stack”.

It is too early to tell whether the positive changes wrought by AI will outweigh the perils. But leading a company in the years ahead is sure to be more challenging than at any time in living memory. AI will require bosses to rethink how they structure departments, whether they should build strategic technologies internally or trust outside firms to deliver them, whether they can attract the technical talent they need, what they owe their employees and how they should balance their strategic interests with workers’ privacy. Just as the internet felled some bosses, those who do not invest in AI early to ensure they will keep their firm’s competitive edge will flounder.

Janus, the Roman god, contained both beginnings and endings within him. That duality characterises AI, too. It will put an end to traditional ways of doing things and start a new era for business and for the world at large. It will be pervasive, devastating and exhilarating all at the same time. Look ahead.