Do Quality Assurance testing models fall short when it comes to Artificial Intelligence / Machine Learning software?

It's a human factor. In my mind, you can't have a machine validate a machine because you'll run into the same issue. You need to have that diversified, unbiased view asking questions: is this leaning towards one way, or the other? What if? Why? And that is a human factor. 20 or even 40 years from now the only job that we’ll have is validating whether or not the output of all of these automated systems is just, ethical, correct, or otherwise. It's a very tough question.

Anonymous Author
It's a human factor. In my mind, you can't have a machine validate a machine because you'll run into the same issue. You need to have that diversified, unbiased view asking questions: is this leaning towards one way, or the other? What if? Why? And that is a human factor. 20 or even 40 years from now the only job that we’ll have is validating whether or not the output of all of these automated systems is just, ethical, correct, or otherwise. It's a very tough question.
2 upvotes
Anonymous Author
The heart of this journey stems from two dynamic places of origin—one, is the approach to how we develop our IT professionals and what we consider strong engineering practices in the design and testing of code; and the second is our approach to developing quality assurance professionals when we take a look at the core foundations of determining if a system is fit for use. Now, when you look at those two practices, they are extremely different schools of thought but the operational tasks behind them are actually the same tasks you're testing. You're testing to determine if the user experience/user story is operating as designed, whereas the quality assurance professional is testing to ensure levels of quality that potentially map to ethical usage, especially if they're in a good practice (GxP) space. As technology becomes more intelligent, we as humans will be the ones checking that it's operating the way we thought it would.
3 upvotes