The bots are coming for your job

• When it comes to Artificial Intelligence, no aspect has been discussed more than the future of work
• Three new reports offer key insights on the implications of AI on the future of work. Would you trust a robot over your boss?

A recent report from the California-based think tank Institute for the Future and Dell Technologies, titled Future Of Connected Living, describes how the gap between humans and machines is shrinking. "Over the next decade...everything around us will become more intelligent, communicative and connected," says the report, forecasting that a new dynamic between humans and machines will "transform our lives in 2030".

When it comes to Artificial Intelligence (AI), no aspect has been discussed more than the future of work. What sort of implications will AI have on jobs and the labour market? Will machines give human employees more time or will they take over their roles altogether? Will AI cause a major disruption?

These have been the key concerns in some of the AI reports and surveys released recently. A report by Mark Muro, Jacob Whiton and Robert Maxim at the Brookings Institution's Metropolitan Policy Program says AI could affect work in virtually "every occupational group". Others look at the way organizations use AI, the problems they face in adoption, and even whether robots inspire greater trust than managers.

The report, What Jobs Are Affected By AI? uses a "new way" to identify the kind of tasks and occupations likely to be affected by AI's machine-learning capabilities, "rather than automation's robotics and software impacts on the economy". The method, proposed by Michael Webb, a PhD candidate from Stanford University's department of economics, predicts the impact of technology on different occupations. In the report, job exposure levels to AI were established by analysing the overlap between AI-related patents and roles in the workplace.

Webb explains in his paper (The Impact Of Artificial Intelligence On The Labor Market), written earlier this month, that the texts of patents contain information on technologies, while job descriptions contain information about the tasks people are expected to perform. Webb uses "verb-noun pairs" to quantify the "overlap" between patents and tasks.

He explains: "Suppose a doctor's job description includes the task 'diagnose patient's condition'. I use a natural language-processing algorithm to extract the verb-noun pairs from this task, which in this case would be 'diagnose condition'. I then quantify how many patents corresponding to a given technology contain similar verb-noun pairs, such as 'diagnose disease'. I use the prevalence of such patents to assign a score to the task, and aggregate these task-level scores to the occupation level."

The Brookings report says better-paid, white-collar occupations may be most exposed to AI, along with agriculture and manufacturing positions. A related finding suggests that prime-age workers (aged 25-54), mostly men, white and Asian Americans, will be affected by AI. Although this report focuses only on the US market, it still offers indications on how AI is likely to impact work in other markets.

Newsroom bot

A report by Polis, the journalism think tank at the London School of Economics, surveyed 71 news organizations across 32 countries (including India), and found that while AI is already giving journalists more power, it comes with its own set of challenges.

The respondents included newspapers, broadcasters, publishing groups, news agencies and magazines.

For the survey, journalists already working with AI responded to questions about their understanding of the technology and how it was being used in newsrooms. They were also asked about their views on potential risks for the news industry. The report (New powers, New Responsibilities: A global survey Of journalism And Artificial Intelligence) describes these respondents as "digital early adopters"—people who had more awareness of AI than a random sample of journalists.

Newsrooms today are using AI broadly in three areas: news production (content creation, editing, packaging of different formats and platforms), news distribution (marketing, finding audiences, monetization/subscriptions, etc.) and news gathering (sourcing of information, generating story ideas, etc.).

The Washington Post, for example, adopted a robo-reporting tool, Heliograf, for the 2016 summer Olympics and congressional races on Election Day that year to churn out hundreds of stories. Bloomberg has an automated system that uses AI to identify key data points in the earnings reports of thousands of companies and "publish headlines and articles in seconds". That's not all. A recent story in The New Yorker by John Seabrook looked at how predictive-text technology could transform the future of writing. Co-writing the story with Seabrook was GPT-2, an AI-writer developed by the company OpenAI.

According to the Polis report, 68% of the respondents said they started adopting AI technologies for work efficiency, 45% said they wanted to deliver more relevant content to users, while 20% said they wanted to improve business efficiency.

This doesn't mean, however, that news organizations have AI strategies in place. The report says just over a third (37%) of the respondents claimed to have an active AI strategy, while nearly two-thirds (63%) did not.

Respondents cited the lack of financial resources (27%) and AI-related skills, along with the difficulty in attracting or hiring talent (24%), as some of the biggest problems in the way of AI adoption. "But as significant as either of those was cultural resistance (24%), including the fear of losing jobs, of changing work habits, and a general hostility to new technology. Lack of knowledge about AI (19%) across the news organisation along with a lack of strategic managerial insight (17%) were also key issues," the report adds.

A third report released this month, based on the AI At Workstudy conducted by technology company Oracle and HR advisory and research firm Future Workplace says that some people trust a robot more than their manager or boss. The study of 8,370 employees, managers and HR leaders was conducted across 10 countries. The findings show how AI is changing the relationship between technology and employees at work: 64% of people worldwide would trust a robot more than their manager; this number rose to 89% in India. Meanwhile, 82% of the respondents think robots can be better at things than their managers. The survey respondents said robots were better at providing "unbiased information", "maintaining work schedules", "problem solving" and "managing a budget".

AI, clearly, is here to stay.