The jobs AI won’t take from us

“I don’t trust these ATM machines. I always withdraw my money at the bank employee’s station. I want my money to be handled by humans, not by robots”, says the strange old lady standing in front of me in the bank queue. Her statement makes me nervous. Apparently what I consider an efficient tool, she sees as a horrifying technical invention. But something tells me I might be standing behind my future self.

Automation has already begun

ATMs are not AIs. But they are a metaphor for the ways technology subsitutes human workforce. Take for example ticket issuing, check-in and luggage drop off at the airport: Automation has eroded its way through numerous industries into our daily life- steady, slowly and unnoticed. With it, thousands of jobs have been sacrificed for the sake of speed and cost reduction. In fact, automation means economic efficiency.

Admittedly, ATMs do not need breaks. They don’t want to leave at 5pm sharp to return to their awaiting families. On the contrary, they can carry out a task in a never ending, tireless and cheap way. Sure, after their introduction to the market it took a while for the public to adapt. But with every generation that streamed pass ATMs, the disturbance, the mistrust and the complaints did ebb away. Today, hardly anyone would doubt the trustworthiness of an ATM and demand from his or her Bank to hire an employee for cash dispensing.

So what if science told you that within the next 20 years, 45% of all known jobs will disappear or be upgraded in a way that human workforce becomes obsolete?

Who’s next?

In 2013, Oxford University calculated exactly this likelihood and published its results for 700 different jobs in total. Ever since, the findings have been a matter of vivid academic discussion as well as massive public disturbance. Why? Because it is hard to admit that the job a human has spent years to master can be handled by an algorithm not only easier, but also better. But it can. And it will.

With a 99% percent likelihood, library and mathematical technicians, and telemarketers will find themselves unemployed due to their more efficient AI successors. Drivers, cashiers, travel agents and fast food workers are next- amongst others. Amazon Go for instance, Amazon’s “new kind of store featuring the world’s most advanced shopping technology” is a perfect example that shops do not depend on cashiers at all. Its claim “no lines, no checkout – just grab and go!” indeed sounds like a blessing for the fast, non- stop 24/7/365 U.S. economy.

Arts and Healthcare won’t be spared, either. At this very second, AIs compose poetry and classical music that is hard to discern from their humanly created competitors. Moreover, there is a prevalent tendency towards the digital storage of medical data. This data could easily be made accesible to algorithms specifically trained to diagnose diseases and suggest treatment accordingly. This opens up a disscusion about  who to trust: A burned-out human physician or a precise machine which is aware of all of your medical history?  Babylon Health “is helping to solve an increasing range of healthcare challenges with artificial intelligence” already. On the other hand, a machine cannot see if your skin is pale or your eyes are glassy. At least not yet. In comparison to the driving or travel agent industry, the Healthcare sector will not be invaded by artificial intelligence anytime soon.

The jobs AI will not take away from us

Hence, occupations with a high level of specialization will be the first to become automated. On the other side, jobs dealing with complex social situations are considered “safe”. According to the Oxford study, these are the positions with the lowest (0.35 or less) probability to disappear within the next decades:

  • Recreational Therapists
  • First-Line Supervisors of Mechanics, Installers, and Repairers
  • Emergency Management Directors
  • Mental Health and Substance Abuse Social Workers
  • Audiologists
  • Occupational Therapists
  • Orthotists and Prosthetists
  • Healthcare Social Workers

In his TED Talk, computer scientist Kai-Fu Lee suggests that artificial intelligence might take a very large number of occupations away from people, but will leave room for more human jobs such as teaching, caretaking and therapy. Is automation the panacea that will free us from our routine jobs and reconnect us to our human values? Will we be able to pass dull and boring jobs on to machines and start building a more compassionate and peaceful society?

Given that the future employment market will strongly rely on occupations dealing with complex social problems, high inequality and poverty rates will be crucial for the maintanance of all these new empathety driven jobs.Thus, the higher the levels of drug abuse, homelessness and social marginalisation, the higher the employment rate for all of us who lost their jobs due to automation. In a nutshell: A demand of social workers and therapists must be created to satisfy this enormous offer of workforce in the future. Without social problems, no social work jobs, no salary. 

Building on this argument, AI will not save humanity by building an empathy utopia. On the contrary: It will trigger an unprecedented inequality crisis to keep a minority employed and wealthy. The alternative? To accept unemployment and invent something new to pass all that free time while also making a living.

But before we theoretically chose between a robot- controlled dystopia for the many and an empathy paradise for the few, some practical questions need to be answered first.

  • Will there be an ethical codex for AI, and if so, who will compose it? Will it be applicable on a national or on an international level?
  • How to patent work created by algorithms? What legal implications are there?
  • What will happen to all the unemployed workforce, is there a plan for their insurance and pensions?
  • Is the eventual goal of AI technology to make life better? And if so, what does better mean?

In case you are interested in further information, we suggest Homo Deus by Yuval Noah Harari, as well as Sam Harris’ “Can we build AI without losing control over it” and Nick Bostrom’s “What happens when our computers get smarter than we are?”

picture credits: Ian Espionosa via

Share this post

Leave a comment

Your email address will not be published. Required fields are marked *