0
$\begingroup$

Say a company in my future world has an algorithm that can evaluate and hire job candidates? Would this potentially cause problems like bias or hire people unfairly? Could it also be unreliable?

$\endgroup$
6
  • $\begingroup$ If that's what your story require, why not? Can you provide some details? $\endgroup$
    – Mołot
    Oct 8, 2017 at 18:43
  • 2
    $\begingroup$ It would be inaccurate if they choose an inaccurate algorithm. We need to know more about the algorithm if we're to tell the chances of the algorithm being inaccurate. $\endgroup$
    – Geo
    Oct 8, 2017 at 18:51
  • $\begingroup$ We need more details about how your algorithm works. As first instance It seems reasonable that the hiring system could be hacked. $\endgroup$
    – user489150
    Oct 8, 2017 at 18:54
  • 1
    $\begingroup$ I don't know about your workplace, but where I work they do have a procedure for evaluating and hiring job candidates. Most of the time it works reasonably well, but occasionally it makes mistakes. I think that such a situation is common, it was common in the dark past when the first employer hired the first employee to work in the mammoth hunting department, and it will be common in the future. P.S. Algorithms are mathematical objects. They cannot be unreliable. It may be the case that the algorithm is misunderstood, misused or not well suited to the problem, but that's different. $\endgroup$
    – AlexP
    Oct 8, 2017 at 19:01
  • $\begingroup$ It really depends on how the code is written. I've put this on hold as, without those additional details, we can't really answer this -- you can make the answer whatever you want it to be. Please edit in more details and the community will review. Thanks. $\endgroup$ Oct 8, 2017 at 19:35

1 Answer 1

0
$\begingroup$

Potentially yes

It certainly could, depending on how it had been programed. Computer code is normally tested before use especially if it is to be used in an important application. However no testing is fool proof and often errors or bugs do not manifest themselves until later when the code is in use. There is an old saying in programming circles that the trouble with making something fool proof is that fools are so ingenious.

It is also possible that it would not if sufficient care had been taken in the design of the code and sufficient testing had been carried out. Critical systems such as plane autopilots and similar often have to undergo particularly gruelling testing before going into operation to prevent or minimise errors. It really depends on the potential problems that might result.

In your example one way to test it would be to run the algorithm in parallel with human reviewers for a while and compare the findings of each.

$\endgroup$

Not the answer you're looking for? Browse other questions tagged .