We have all been talking about itā¦ But how much do you agree with this statement: "AI will replace the majority of testing roles by 2040?
bonus bots for your thoughts
- Strongly disagree
- Somewhat disagree
- Somewhat agree
- Strongly agree
0
voters
4 Likes
Iād say AI will transform software testing for sure (a lot).
3 Likes
I donāt feel AI will replace a majority of testers.
However, I certainly feel that AI will replace a majority of āpseudoā testers.
Testers who are probably not learning the craft or just doing basic script writing work or not adding value beyond some base-level work.
Serious testers are and will always be high in demand.
Testers who can leverage AI and build intelligent systems will also be in top demand.
The real testing experts will be more valued as they can be the captain of this AI-driven ship 
9 Likes
Its another tool to help you test.
Just other day, I used chromeās dev tools recorder feature to create an entity of which I needed 10 more. I then fed that to an LLM and it gave me a JS code to run in the console and voila! I got my test data.
P.S couldnāt do it in another way because they didnāt expose any other method for this.
1 Like
No itās not going to be replaced, for example; automation with code / no-code came, giving more helping hands to do work efficiently and quickly.
Similarly, QA, which understands and knows how to use it, can provide speed, accuracy, and more coverage.
Monotonous work reduces for QA & they can focus more on other QA activities.
1 Like
Its worth a consideration of your current model of testing.
There are a lot of elements of testing that favour machine strengths sometimes some companies have their whole model of testing that favours these strengths but have humans still doing the a fair amount of the testing. So here it can be having humans working to machine strength areas.
So a significant part of the replacement question is whether the majority of testing roles fit the above model and if they do, personally Iād see a higher transition to AI.
Another model is having some testers working more towards their very human strengths, things like applying curiosity, exploration, discovery, learning, investigation, experimentation, empathy, collaboration, context awareness, variation and a whole load of skepticism alongside a big toolbox to their testing.
So the idea of leveraging machines and AI for areas that favour their own strengths and utilise the wonderful human traits for the tasks they are better suited for, is likely a balance that should be there for a long time.
There is a risk though that the majority of testing roles currently sit in the machine strength area and not a balanced one that consciously leverages from very human strengths.
2 Likes
Generative AI is just a tool. If you believe your job can be replaced by a tool, it might be time to reassess your role and skills.
(Comment generated with the help of Gen AI :))
4 Likes
My initial expectation when seeing this question, was to see simular responses as āwill automated testing replace manual testingā
And my feelings are the same, AI should enhance our testing, maybe make us more efficient, but wonāt replace testing roles.
2 Likes
Guessing about anything 2040 is beyond meā¦
2 Likes
Itās hard to say what will happen till 2040. Looking at the current speed of AI, the whole IT industry is changing dramatically. And I canāt even imagine what the IT industry will look like in 16 years. However, I am confident that you always need humans to validate the results of AI. Our roles will change, I am looking forward to the change 
1 Like
Iām not a big fan of the term āAIā. True artificial intelligence is a very different subject. What most people are discussing is machine learning, which has, for some reason been referred to as AIā¦ (mini rant over)ā¦
Current machine learning or AI as itās known 
, has shown time and time again that the responses returned can not always be trusted whether it be generally inaccurate information, a response that doesnāt actually solve the question asked, other aspects like bias or even a malicious action/output.
A human end-user of such systems can take the information returned and judge if this is helpful to them and if it will provide the correct end result.
AI/ML may work well for a while, but companies that rely solely on āAIā to ensure that their software is fully tested, will likely come unstuck at some point (in my opinion).
Complex tech like āgenerative AIā can and will be manipulated and hacked, there are already examples of this happening. (See attached articles). So going forward, there will likely be a need for humans to ensure that the outputs received are not only correct but also safe to use.
And yes, AI / ML may get better and better as time goes by, but you will never stop the hackers!
1 Like
AI is definitely changing software testing and software development as an effective tool. There are other factors that lead to significant changes in approaches to QA and testing in many companies. But modern LLM-based AI has its limitations, currently, it seems it wonāt be able to become a real AI to replace the majority of testing/QA/dev roles but itās a matter of prediction.
To answer the question you need to understand how modern AI works and if there is any potential for replacing people or if it always will be more like an effective tool that will change many roles but make them more efficient.
Iām not an expert in this field as the majority of people here, and even real experts avoid any serious predictions about AI. From the tendencies of developing modern tech and how fast AI tech developing we assume that the same trend in the development and improvement of AI will continue - the same progression but as we can see with science and other technology the real progress of a particular tech quite often slow down and even stopping to the moment when we just use it as is in a different context but not really developing the tech by itself.
Check some videos about AI on this channel if you want to get a more scientific point of view and reasonable logical skepticism about its future https://www.youtube.com/@SabineHossenfelder/search?query=AI
LLMs are another bubble waiting to burst. Mostly it is hype built up around a trend that will ultimately fade.
The biggest thing about them is that you cannot rely on the answers being correct. Essentially, you need to know the answer yourself, before you ask a question - so that you are able to validate the answer given. This reduces them from oracles to second guesses. They can be used to iron out errors caused by flawed humans that should know better. They are an amalgam of knowledge, not a source of it.
LLMs have their uses, but they are far from the silver bullets that their hype would have us believe.
AI will NOT replace testing roles.
How I see it:
"Testers will be replaced by Testers who use AI. "
And this is the reason I picked āsomewhat Agreeā
2 Likes
2040? Better ask for 2025 
1 Like
2040 is a long way off, 16 years in Technology is a huge time. I honestly donāt know what the tech world will look like in that time. I have been in this industry for 30 years and it is completely different from when I started. That was without the advance of AI and truly accelerated path that we are on.
I think it will come down to trust, from us and from our customers. As AI develops it will drastially change our landscape. I think it will require much more skilled Testers, developers, Quality Engineers. I think as we āengineerā the humans out of the creation of our products it may become more important to ensure that we add more of the human factor into the testing.
Whatever happens itās going to be an interesting ride?!
1 Like
For me, I see the conversation about AI as the same as the ones weāve had about automation, Agile and offshoring. In each of these cases, something new was adopted by some of the industry and has a short term effect on testing but then balanced out.
Yes, some organisations started wanting different skills or maybe slightly fewer testers. But in the most part there were still lots of different testing roles out there using existing and new skills - thatās why even today there are roles available for scripted manual testing and waterfall, they didnāt disappear.
Some organisations might move to low code / no code testing and code generation, then bounce back to how things are or maybe embrace it. Either way thereāll be jobs for those that adopt this technology (and can maintain it) as well as testers with other focuses.
1 Like