We continue to see the doomsday narrative: “Testing as we know it today is dead. Humans will be replaced by artificial intelligence in the near future.” But AI is not here to replace human testers. Looking into a crystal ball, testing still looks promising for real people.
The trick is that humans will need to learn to coexist with these testing bots. Here are some ways we can start to think differently about the future of testing.
Finding bugs: It’s true that testers can work for about eight hours a day, max, while a QA bot will work tirelessly 24/7 without interrupting production. But after the bots found all those bugs overnight, the next morning humans will need to triage and resolve them—or maybe run the future systems that will automatically rewrite the code.
Increasing test coverage: When a product expands, instead of cutting corners for what should be tested in a development cycle due to the limited number of people on the testing team, those team members can use QA bots to autotune parallel testing efforts. And if testers find more work, they can clone an army of QA bots to cover those untested areas.
Performing real-time analysis: Analysis of real-time data will require a tester to perform a real-time stream of QA for that data. Again, humans can get assistance by using a QA bot to clean polluted data and report deviations.
There’s a pattern here: Human testers aren’t going away; they will just need to figure out how to collaborate with their nonhuman counterparts for an optimal QA effort. Although artificially intelligent bots likely will be used in the future due to their speed, continuous operation, and low error rate, testers will still be needed to design tests, tell these bots what to do, and attach real meaning to the results.
There will be an initial investment to set up an AI system, but once that is done, organizations will get more testing for less money. That means they can use that savings to invest in more QA innovation efforts, such as figuring out how to test uncovered areas and devoting more time to exploratory testing.
I think the way to go is to retrain humans into a QA ops type of profession, with a focus on monitoring QA bots, QA specialists, and their results, and then support their efforts for contributing to even better quality.
In the long run, artificial intelligence will help not only the QA engineers who want to solve the challenge of getting higher quality products to market, but all roles involved in software product development. Testers don’t need to worry about their careers becoming obsolete. They just need to start thinking differently about the future of testing.
Alexander Andelkovic is presenting the sessions Using Artificial Intelligence to Test the Candy Crush Saga Game and Internet of Fun: Winning Ways for an IoT Hackathon at the IoT Dev + Test 2017 conference, April 24–28 in San Diego, CA.
