Is Your Job Up For Grabs By Chatbots?
By Adam Blair, Executive Editor
Artificial intelligence (AI) is threatening/promising to remake the world’s workforce, and the retail industry will not be spared. (Nor will journalism, for that matter.) The futurists, analysts, brand and CPG executives who spoke at the ai.now workshop sponsored by Fractal Analytics, held July 18 in New York City, shared an over-arching warning: If your workload is data-driven and/or highly repetitive, a combination of AI and robotics could make your job obsolete.
One of the simplest examples is in health care: Nurses and primary care physicians who interact with patients are more likely to stay employed compared to radiologists, who examine X-rays for tumors and other health issues. Image recognition solutions powered by machine learning can be trained to review these X-rays, and they can examine many more than a human doctor ever could — and in some cases with greater accuracy.
The future isn’t exactly the bleak dystopia of the Terminator movies, however, even if the keynote by futurist Martin Ford was titled Rise of the Robots. Ford, author of a book of the same name, provided data that revealed how much the world of work has already changed due to changes in information technology. “New industries are not labor intensive,” said Ford, making the following comparison:
• In 1979, General Motors generated $11 billion in earnings (measured in 2012 dollars) and employed 840,000 workers;
• In 2012, Google generated $14 billion in earnings, 20% more than G.M., but employed only 38,000 workers — 4.5% of G.M.’s workforce at its peak.
“Since the 1960s, each succeeding decade has produced fewer new jobs (by percentage),” Ford noted. “There has been a hollowing out of the job market, and that’s almost certainly a result of technology.”
White-collar jobs are not exempt from these trends, Ford added: “In 2004, the median staffing of a corporate finance department was 120 full-time employees for each $1 billion in revenue,” he said. “By 2014 it was down to 71.26 employees. In journalism, the development of smart algorithms can tap into a stream of data, find the most interesting elements and create an article. One of these is published approximately every 30 seconds.”
People Skills Will Save People’s Jobs
Retail jobs that require complex human interactions — a sales associates helping a customer find the right product, for example — are less likely to be usurped by AI-powered technologies. “Human-to-human interaction is still a critical skill,” said Matthew Keylock, Head of Data and Analytics for Connected Solutions at Mars Petcare. He added that AI solutions rely on data, and lots of it, to function, and “there’s so little of context that can be captured in a data form, so understanding relationships and their context will still be important.”
People who can understand what’s happening inside the “black box” of an AI-powered solution also will remain employable. “With AI, governance is key,” said Keylock, noting that existing cognitive biases can affect how an AI/machine learning solution develops. “Companies need not only technical experts but also business translators, and those can be harder to find than the technical people. You don’t want an [AI solution] to run riot where you don’t know what’s happening with it minute to minute.”
How AI And Humans Can Work Together
Many AI-based solutions can serve as tools that improve a human’s ability to do his or her job. For example, “If you’re a typical customer service rep, you’re not set up well to succeed,” said Kjell Carlsson, a Senior Analyst at Forrester. “The economics of call centers means you aren’t paid very much, and you have no information about the person calling in or the types of things they are calling about.”
With an AI-powered “listening agent” capable of voice recognition and sentiment analysis monitoring the calls coming in, however, “the solution could present 10 likely options for what the call is about,” said Carlsson. “Then the human can quickly identify the one that is likeliest.”
Companies may be tempted to do a full rip-and-replace of their customer service apparatus and replace humans with chatbots, “but if you start with the perspective of augmenting someone that is already doing this job, it’s easier to make headway,” Carlsson noted. “If you can make a person just slightly more efficient, you can measure that. However, the sexiness of full automation and a soothing voice coming from the black box are the AI applications that get all the attention.”
One final note: This article was not written by a robot — but if it was, would you be certain that you could tell the difference?