“Testing is dead.”
“AI bots will replace us in testing jobs.”
“The end is near!”
Have you heard fellow testers say these things? I have—some for almost forty years. As Mark Twain once said, “The reports of my death are greatly exaggerated.”
Even though I have seen many changes throughout my software and testing career, I do believe AI will bring change that, if you are not ready for it, may limit your future.
I have supported AI and played with it in the past, with interesting results. It has great power that must be respected and understood. The AI bot I taught learned that code metrics are not much use as an indicator of quality, but that programmers who write more comments (not necessarily better comments, just more of them) make fewer coding errors. It may sound strange, but this was backed up by other research I have read.
Here are some other impending changes due to AI that I think you may want to consider and work with.
Update Your Skills
First, if you are a manual tester, running tests mostly by hand and following written test procedures with little or no thought, you should be afraid. Test automation supported by AI will likely replace you. This is true of any job in any industry where there are repetitive tasks that a robot or computer can do faster than you can, and cheaper, too.
It has been happening for years, so you should take actions now to start updating your knowledge and skills.
One skill to update is creative thinking demonstrated through exploratory testing. Exploratory testing is something humans still excel at.
Next, if you are doing some kind of test automation, you need to be paying attention to AI and AI-supported test automation tools. These tools are evolving, and you need to understand this integration change. It is likely the AI bots will take over this generation, and then execution will happen using automated test tools running hundreds, thousands, and even millions of tests. If you are not doing test automation but are thinking about it, you should pay attention to AI automation trends.
Additionally, if you do not think about things like test management, planning, strategy, architecture, and improvement, you are missing a major boat that will set sail on the AI winds without you. If these terms are unknown to you, you have another learning assignment.
The time is now to do some reading and research on these topics. Learn new things and build added value in your skills. I have spent my software testing life constantly learning and building new skills. If you are not learning new things, you will become a dinosaur quickly in the software world.
Evolve with the Technology
Although these changes driven by AI may seem intimidating, the good news is that I think we may all get to work on more fun things—and possibly also have more days off. Who would not want to work three or four days a week, monitor what the AI system is doing while we play, and get better software, all at the same time?
Software is going to be everywhere and in everything. This means we will need a lot of quality testing done by highly skilled, critical-thinking engineers.
I teach people around the world who want the easy answer and a “silver bullet” for testing. They even get mad when I tell them they must learn to think and grow. These types of people are the ones who are endangered.
Agile was going to be the silver bullet. Automation of testing was the easy answer. Neither was totally perfect, easy, or any kind of panacea. Now it is AI’s turn to save the industry. Ha!
I believe we must change with technology’s evolution. I am no longer using punch cards, and I now program parts of my life by talking to my phone. Times change. We must change with them.
Am I an optimist? Sure, but in the last forty years, my optimism has proven right more often than not. Who wants boring “follow-the-script” manual test jobs?
In terms of test planning, strategy, and upfront architecture, I personally gravitate toward model-based testing to drive test automation that is interfaced with AI. In these test plans, my strategy will be to create complex models in languages such as Unified Modeling Language Testing Profile (UTP) that can be expanded, reused, changed, and refactored to generate many tests within a smart, model-driven test environment.
I have tested critical software systems where the testing was driven by keywords and models that functioned in real time to allow high degrees of automation while collecting terabytes of test information. A tester would start a system test that would run thousands of automated test cases over the software under test in real time. During this time the tester could enjoy thinking about new models and strategies.
The use of planning, strategy, and model architectures will not be the only answers when dealing with the AI world, but they are the ones I am comfortable with and confident in.
Don’t Fear the Future
Elon Musk says we should be worried about the AI threat. I say you can be better than the AI, but you should respect it.
Some people say that AI will take over jobs in the future. I say, cool—most of those jobs were boring anyway, and I enjoy creative, challenging tasks that AI seems to be bad at.
Many users complain about smart AI devices and their lack of safety. I say it sounds like a job for testers using AI automation to me. I wouldn’t mind that three-days-a-week job hacking AI device security.
Mostly, I say that with the right mindset and the right skills, you can have fun in the robotic AI future—and maybe even work less and play more.
User Comments
Such an interesting article, Jon. And, I really like your humor. I completely agree with you. Instead of fearing AI, we should make it a friend who lets us have an extended weekend, every week. Of course, that is utopian but not exactly impossible. We humans can limit ourselves to creative and supervisory tasks.