18 Mar
When thinking about Artificial Intelligence (AI), most people think of “automated control”. A “Big Brother” that takes over our jobs and our lifes, radically changing the way we are used to live (just think of scenarios pictured by movies such as “I, Robot” or “Terminator”.
Artificial Intelligence (AI) has indeed revolutionized a wide range of areas in the last years, from driving and healthcare to customer interaction and human resource management.
However, this was also the case for the internet, just a few years ago. Or the telephone, some more years ago. Just to name two examples…
I mean phrases like AI-based software testing, or even AI-based software development actually sound very Sci-Fi (e.g. think of AI developing itself?).
There is also a number of articles out there that try to claim that due to these phrases becoming reality bit by bit over the next few years, developers won’t be needed in the future.
So, let me tell you why AI will never actually take our jobs (and stop telling me that it does):
- AI is just an assistant. The problems that AI can solve today are very small, compared to the tasks that humans perform every day. Just think of one big breakthrough in AI (e.g. distinguishing cars from people), and then think of whether you ever cared about performing this task yourself. In general, AI is not really capable of performing the tasks that actually takes cognitive effort from us. And even if it could, the problem is that AI works (i) based on probabilities, and (ii) most often based on historical data. (i) Working based on probabilities means that you can never fully trust a result that AI produces – it is just the most probable answer to your question! (i) Working based on historical data means that it is hard to explore “new” things in the sense of actually being creative – which is where humans tend to thrive!
- We need AI to be more productive. We are usually not interested in doing small, repetitive tasks. However, these tasks still take most of the time of most people’s workday. But these repetitive tasks are also where AI (and particularly Machine Learning) is most efficient. Why? Because there is a lot of those tasks. So let us use AI to help us become more productive!
- Being more productive means creating more, not working less. By being more productive, we can start exploring new things that we would usually not have the time to. Life offers us a great deal of challenges which we usually don’t have the time to explore – this is why we need AI to help us become more productive! Of course, we could also use this increased productivity to simply work less, giving us more spare time, but eventually also increasing unemployment rate (less work = less job positions). However, history shows us that humans tend to find new ways to keep themselves busy. E.g. think of the tremendous impact that the internet had on our productivity, when it comes to communication or information sharing. But as a response, people did not work less, because information and communication did cost way less time – they started creating even more things. This is how economies grow!
- There will always be something new to develop. Many people still fear that if they can’t do these repetitive tasks most of their day, they will have nothing left to work on. But let’s be honest: as a developer, did you ever fear that there will ever be nothing you can work on? If you do, just send me a mail and I will instantly give you 5 great project ideas that I wanted to work on over the last years, but simply did not have the time to. Or just go on Github and do some Open-Source contributions. There are still so many things out there that we can still explore – so many problems, that we can solve. Problems that people are struggling with each and every day. But to do that, we need the time to build stuff. And (i) making developers more productive, as well as (ii) shifting labour from repetitive work places to creative ones will help us in providing the necessary workforce. So let us use AI to make us more productive, create more output, and make a better tomorrow!
The above mentioned trend is is also supported by a study conducted by Gartner, that concluded that by 2025, AI will already bring 2 million more jobs than it will take, worldwide. A summary of this study is available here: gartner.com and you can also download the full report here: gartner.com.
But that was my opinion – what do you think of Artificial Intelligence, and the future of software development (software testing in particular)?
Let’s discuss in the comments section below!
written by Daniel Lehner
Abstract vector created by vectorjuice – www.freepik.com
Daniel Lehner is a PhD student at the Johannes Keppler University in Linz, Austria. His research focus is on traceability of software information between different stages in the development cycle. During his master studies in Business Informatics at the Vienna University of Technology, Daniel worked as a consultant in the area of software quality assurance.
Information
Daniel Lehner