Artificial intelligence continues to be a rapidly developing field with fascinating new advancements. For example, a team from Botnik Studios fed a predictive text keyboard all seven Harry Potter novels by J.K Rowling, and the system then wrote its own version of a Harry Potter book.
As fun as these technologies can be, many people are weary of one particular faction of artificial intelligence known as machine learning technology, where computer systems improve with more experience.
There is growing concern that machine learning could potentially take over human jobs, as it has already significantly improved facial recognition technology, credit card fraud detection, and financial market analysis.
These concerns are not necessarily unfounded, as machine-learning computer systems could have as drastic an impact on the economy as electricity did when it was first introduced.
But according to two researchers, Tom Mitchell from Carnegie Mellon University and Erik Brynjolfsson from the Massachusetts Institute of Technology, machine learning (ML) will not take over all jobs or replace the workforce entirely.
Mitchell and Brynjolfsson co-authored a Policy Forum which will be published in the journal Science discussing 21 criteria for evaluating whether a job is likely to be completed by ML.
For example, ML is even predicted to be able to detect cancers and conduct medical diagnosis in the future. But does this mean computer systems will replace medical professionals?
Tom Mitchell and Erik Brynjolfsson say that’s not likely to happen, but instead what will change is the nature of jobs themselves.
“We don’t know how all of this will play out,” said Mitchell. “People whose jobs involve human-to-human interaction are going to be more valuable because they can’t be automated.”
According to the authors, ML is perfect for tasks that require sifting through lots of data like scheduling.
But ML would not work with making decisions that require lots of thinking, or giving detailed explanations.
For example, a cancer detection program can study hundreds of thousands of skin lesions and tumors which helps it make an accurate diagnosis, but only a dermatologist or doctor can go into detail as to why a lesion is cancerous.
Time will tell how ML shapes the economy and jobs, and the researchers note that caution is needed when developing ML technology.
“The potential for large and rapid changes due to ML, in many cases within a decade, suggests that the economic effects may be highly disruptive, creating both winners and losers,” write Mitchell and Brynjolfsson. “This will require considerable attention among policymakers, business leaders, technologists and researchers.”