Could A Robot Take Your Job?

Are break-throughs in Artificial Intelligence about to consign fellow humans into employment wilderness?


The BBC in the UK recently published a study of over 350 professions, ranking them in order of the likelihood of a robot being able to do their job in the near future. The data, gathered by Oxford University academics Michael Osborne and Carl Frey, judged each position against 9 key parameters of artificial intelligence and machine learning in order to compile the rankings. The parameters against which each position were judged were:

  • Social perceptiveness
  • Negotiation
  • Persuasion
  • Assisting and caring for others
  • Originality
  • Fine arts
  • Finger dexterity
  • Manual dexterity
  • The need to work in a cramped work space

It would appear as though these parameters have been very carefully chosen, as they reflect the main differences between a human worker and their artificial intelligence equivalent. In particular, areas such as originality, assisting and caring for others and social perceptiveness are where artificial intelligence struggle due to their inability to register or convey emotion. However, should we be preparing for a future in which robots can do the jobs of humans more quickly, safely and arguably most importantly, less expensively? With that question in mind; we've compiled a list of five professions that might be replaced by a robot in the not-too-distant future.

1. Taxi Driver

Autonomous cars are a very exciting development in the world of artificial intelligence, as evidenced by the news that as of September 2015, Google's fleet of 12 driverless cars have completed over one million road miles, the equivalent to 75 years of adult driving time. While these tests are not completely flawless, there is evidence to suggest that soon, there will be fleets of driverless taxis on the streets of the world's major cities. As early as 2013, Google was requesting the chance to commission the first autonomous taxis. Revolutionary taxi firm Uber are trying to stay ahead of the curve, investing $5.5 million in Carnegie Mellon University's Robotics Department after poaching 40 of their top research scientists, with their long term goal of creating an autonomous fleet of Uber taxis. Once the technology is perfected, it will only be a matter of time until driverless taxis are seen on the corner of city streets all over the world.

Likelihood of Autonomy: 7/10

2. Doctor

Another area in which semi-autonomy has already been achieved, the medical profession is an area in which artificial intelligence development would be welcomed with open arms. A company called iRobot have already produced 'robot doctors' for a small number of hospitals across North America, that are able to bring a top consultant to the bed of a patient over video, with the doctor able to diagnose the patient's symptoms using a range of on-board functions. Away from the hospital, has been heralded as the future of at-home diagnosis, using artificial intelligence algorithms to rule out possible diagnoses using symptoms input by the patient. The most exciting AI medical development will be that doctors and surgeons will have access to a much larger collection of medical data than before, both on the patient they are treating and on previous cases of the conditions, but this doesn't necessarily nullify the use of a human doctor, who might be preferred by patients for their compassion and understanding to a autonomous equivalent.

Likelihood of Autonomy: 6/10

3. Factory Worker

In much the same way the industrial revolution that saw production line workers slowly replaced by machines, it is very likely that in the years to come, humans will have little to no role in the factory setting, with many robots now able to autonomously build the machines required for the production line without any human input. Bar the prohibitive start-up costs, autonomous factory workers are infinitely advantageous to their human counterparts, being more efficient, safer and less prone to striking, and arguably cheaper in the long run when wages and pensions are considered. The only downside to these robots are that they are not able to fix themselves, meaning human technicians would most likely need to be nearby in case of incident, which can potentially cost millions of dollars if not handled with the appropriate care. Overall however, it would appear as though the days of the human-dominated production line could be a thing of the past.

Likelihood of Autonomy: 9/10

4. Telephone Salesperson

Much like the factory worker, it would appear as though technology is advancing too rapidly for the telephone salesperson to be a long-term career choice for much longer. As evidenced by everyday technology such as Apple's 'Siri', Artificial Intelligence is now able to understand, process and respond to even the most obscure human requests. At the more advanced end of the scale, IPSoft's Amelia AI is revolutionising the way software interacts with humans, with its ability to use big data to learn from its mistakes. Although this sounds more like fiction than fact; Spike Jonze's 2013 film 'Her' comes to mind, the technology is very much a reality. But could a technology ever convince a human to buy a product? There is something exclusively human about the ability of a salesperson to convince their client, but is there potential for autonomous software to learn this tactful art? The advantages to the companies are limitless, but foremost in their minds will surely be that you don't have to pay a piece of software commission or a wage, which is likely to be their biggest outgoing in today's ultra-competitive market.

Likelihood of Autonomy: 9/10

5. Psychologist

One area in which it will be almost impossible for a autonomous robot to replace a human would be where compassion and empathy are required. As hard as artificial intelligence researchers try, there isn't an algorithm for sadness, happiness, euphoria or depression, hence why psychology is an area in which humans are superior. This is not for a want of trying however, as several developments, from Affectiva's facial expression-reading software to the Microsoft Kinect analyzing player's bodies during games, but as Meghan Neal of VICE points out, we barely understand our own emotions the majority of the time, so can we really expect machines that we build to understand us any better than we can ourselves? It would seem as though the days of telling all to a robot AI and getting an intelligent, but more importantly, an emotional response might be some way off yet.

Likelihood of Autonomy: 2/10

University lecture small

Read next:

How Are Higher Education Institutions Using Analytics?