“…there are three important factors which constitute a robot: 1) Robots interact with the physical world via sensors and actuators; 2) robots are programmable; and 3) robots are usually autonomous or semi-autonomous.” Alex Owen-Hill (2017) “What's the Difference Between Robotics and Artificial Intelligence”
Just a few years ago talk of robots taking over was all the rage. Thanks to robots, humans would no longer have to work unless they wanted to. Some typical headlines:
When Robots Take All the Work, What'll Be Left for Us to Do? (Wired/2015)
Robots will take over most jobs within 30 years, experts warn (The Telegraph/2016)
Robots will eliminate 6% of all US jobs by 2021, report says (The Guardian, 2016)
Self-driving cars were also just around the corner. Elon Musk predicted cars would achieve “complete autonomy” by 2018. General Motors planned to launch self-driving taxis in San Francisco by 2019. And the boss of Waymo, Google’s self-driving car subsidiary, said in 2015 that he hoped his son, then 11 years old, would never need a driving license.
So what happened? Here are some possibilities:
Real managers in real companies find it difficult to find uses for robots that justify the expense and difficulty of implementation.
Businesses that have made “significant investments” in smart devices and machines are not seeing much bang for their buck. For example, a survey of almost 2,500 bosses found that seven out of ten said their AI projects had generated little impact so far.
Investors have jumped off the smart machine bandwagon as business enthusiasm has cooled.
The performance of robots and self-driving cars has been disappointing. They make dumb mistakes and can get things very wrong if faced with the unexpected.
Because many real-life problems are complicated and require reasoning skills that robots lack, companies have to keep human staff around to deal with difficult cases that robots can’t handle. That limits the expected cost savings.
Smart machines have to be trained on tons of data. A Chinese firm called MBH, for instance, employs more than 300,000 people to label endless pictures of faces, street scenes or medical scans so that they can be processed by machines. That’s expensive.
Well-publicized disasters have cooled the ardor for self-driving cars. In 2018 a self-driving car being tested by Uber, became the first to kill a pedestrian when it hit a woman pushing a bicycle across a road in Arizona.
Semi-autonomous cars are not much an improvement over regular cars. Users of Tesla’s “Autopilot” software must keep their hands on the wheel and their eyes on the road - what’s the appeal in that?
Robots and self-driving cars are unable to cope with unusual circumstances that are not common in the training data. Driving is full of such oddities, such as an escaped horse in the road or a bunch of stickers on a sign. Human drivers usually deal with them without thinking. But machines struggle.
Many job tasks are hard to automatize. According to McKinsey Global Group (2017), these include managing and developing people; applying expertise to decision making, planning, and creative tasks; and operating machinery in unpredictable environments.
There continues to be a market for tasks and services provided by humans. For example, no matter how competent your friendly robot nurse, nothing beats the human touch. Robots may eventually provide some kinds of medical care and treatment, but they are more likely to complement than replace healthcare workers. Just think of them as new members of the healthcare team.
New technologies require qualified personnel to program, monitor, and manage them. That takes time and money. Most businesses just don’t have the resources to jump all in at once. Hence, automation of the workplace will not gain steam until the projected return on investment makes the transition worth it. That takes time.
Much of the previous hype about robots and self-driving cars was based on ignorance of what certain jobs actually entail. For example, truck drivers don’t just drive from point A to point B but have to get in and out of trucks in order perform roadside repairs and adjust loads. If truck drivers have extra free time because their vehicle is self-driving, they may be able to take on additional tasks in the truck cab, such as make sales calls and schedule upcoming deliveries. No job lost but a big gain in productivity.
References:
“An understanding of AI’s limitations is starting to sink in”The Economist , June 11, 2020
“Businesses are finding AI hard to adopt” The Economist, June 11, 2020
“Driverless cars show the limits of today’s AI” The Economist, June 11, 2020.
“For AI, data are harder to come by than you think” The Economist, June 11, 2020.
“The cost of training machines is becoming a problem” The Economist, June 11, 2020.
“The potential and the pitfalls of medical AI” The Economist , June 11, 2020.
A Future That Works: Automation, Employment, and Productivity; McKinsey Global Group January 2017
Arntz, M., T. Gregory and U. Zierahn (2016), The Risk of Automation for Jobs in OECD Countries: A Comparative Analysis. OECD Social, Employment and Migration Working Papers, No. 189, OECD Publishing, Paris.
Levitan, S.A. and Johnson, C.M. The future of work: Does it belong to us or to the robots. Monthly Labor Review; 1982, 105, 10–14.