AI: The Replacement Game

AI: The Replacement Game

Over the past few weeks, amongst a busy work schedule and family life, I've been preparing for a new keynote I'll be giving at different conferences this year entitled: Humanizing AI. I'm following an approach I experimented with last year where I start with a base storyline and have placeholders for dynamic content that I customize for each event so that the resulting presentation is unique for the audience. Clearly, this takes more effort on my part and so to not have the workload build up, I gradually search for the variable or substitute content whenever I have some downtime.

Yesterday, as I was searching on a popular sub-topic for the keynote, the Association for Software Testing made a call for contributions on the exact same subject:

Why don't we replace testers with AI?

This article is a combination of my response to that call, while acknowledging that it seems that as humans we are playing what I'm calling the AI replacement game both in our minds and in reality.

Apparently, we are fascinated with the question:

Why don't we replace [job role] with AI?

Before we get into the discussion about AI and testers, let's plug some dynamic content into the above placeholder to discuss some of the key elements I'll use to shape a point of view. I've included short 5-10 minute videos in each area to stimulate thought and discussion on the AI-related transformation considerations within each application domain.


Banking

Why don't we replace bankers with automated tellers or digital banking?

Because banking involves much more than handling transactions! Bankers offer a personal touch, understand and support customers through significant financial decisions, and provide valuable advice that machines cannot offer. However, this does not mean that there isn't a significant transformational shift here. The key is the tradeoff between the demand and value of human-driven banking experiences and the cost efficiencies and convenience gained from digital banking. Here's a quick look at this balancing act and the future of retail banking.


Teaching

Why don't we replace teachers with online learning platforms?

Because teaching is more than just delivering information! Teachers play a critical role in motivating students, understanding their unique learning needs, and fostering a supportive and interactive classroom environment. As someone who started their career as a college professor, I can definitely relate to the fact that there's a significant gap between engaging in live, in-person instruction versus on-line training, even when human led. The experience is even further removed when the human is taken out of the loop. Here's former middle school teacher and current college professor John Spencer's take on why AI won't replace teachers.

Do you agree with John's point of view? Is teaching really an inherently human endeavor? Feel free to drop some of your comments on this topic below.


Healthcare

Why don't we replace doctors with medical diagnostic software?

Because healthcare is more than just identifying diseases! Doctors provide bedside manner, offer empathy and reassurance, understand the nuances of a patient's symptoms, and can make complex decisions that software cannot. This is not to say that AI isn't playing a major role in the transformation of healthcare. The key here is to recognize that the role and functions of a doctor are highly complex and while robots may takeover some functions in the near future, other functions are much further out and/or may never be replaced. It is therefore better to think about the levels of autonomy when it comes to specific tasks or functions. Here's a clip showing how AI is pushing medical robotics towards autonomy.


Cooking

Why don't we replace chefs with food preparation robots?

Because cooking is an art that involves creativity and passion! Chefs can experiment with flavors, adjust dishes to customer preferences, and present meals in an appealing manner that robots cannot replicate. Okay, so maybe not all cooking is classified as culinary art. Sometimes food is just fuel and unfortunately with our busy lives we sometimes turn towards fast-food. This is where we are seeing a lot more traction with AI. Take a look at the world's first fully autonomous AI-operated fast-food restaurant.


Software Engineering

Why don't we replace software engineers with automated code generators? Because software development is more than writing code! Engineers bring creativity to problem-solving, understand the broader context of applications, and can innovate and design solutions that current automated tools do not. Well at least not yet. By now you may have heard of Cognition Labs Devin AI. If you haven't check out this demo video.

Now seems like a good time to note that although I am really optimistic about what AI can do for our industry, I'm also someone that prides themselves in being able to sift through some of the hype. So will Devin AI take your job and replace you as a software engineer? In the next video, Kyle from Web Dev Simplified breaks down exactly what Devin is, what it can do for you, where it falls short, and why it is unlikely to replace you as a software engineer.


Software Testing

Alright, so it seems like we may have gone half way around the world to finally get to the original question posed by the Association for Software Testing: Why don't we replace testers with AI? Taking some key elements from our discussion so far, here's my point of view:

  • Because testers bring a human perspective to the process, understanding user experiences in a way that AI currently does not.
  • Because testers can think outside of the box to identify new approaches, methods, and techniques for discovering potential risks and issues that AI currently does not.
  • Because testers are needed to validate and verify the AI itself to make sure that learning models are correct, complete, consistent, verifiable, secure, ethical, responsible and more. Until AI autonomy is met with an equal level of trust, there will likely be human-in-the-loop checks and balances.
  • Because AI just isn't there yet and, even if it may be one day, it's success will still depend on how well testers have been able to contribute to making sure that the AI systems being developed meet the expectations of the user.
  • Because artificial intelligence is not human intelligence! Both AI and humans are fallible. At the same time, there are things humans can do better than machines and vice-versa. We're likely to be better off figuring out what each of those things are, the context around them, and then selecting the most appropriate solutions. That is, until something else tells us differently.

During last year's #STARWEST event by TechWell Corp , I ran a tutorial entitled: 'Prompt Engineering for Software Quality Professionals'. After going through a number of prompting best practices, I gave them an exercise to try to guide ChatGPT-4 through prompts to being the best virtual software testing assistant it could be. I haven't shared the results yet as I'd like to run one more instance of the course first and put them together. Thankfully, that will be in a few weeks as I'm running the training at this #STAREAST. But who knows, perhaps it may be possible some day to create the quality engineering version of Devin AI.

Again, not sure it would be up to snuff to replace your testing job, but you never know :) I guess it depends on the kind of tester you are... are you a fast-food tester or a master chef tester?


Wrap Up

It's an exciting, and in some cases nerve-racking, time to be in technology. This is because we see the potential for AI to transform our lives. I firmly believe it will as most technological revolutions do. However, this is not a matter of replacing humans but more likely a matter of transforming certain job functions, hopefully for the better through vetted, reliable automation capabilities. Technology has always shaped our lives and you can expect that trend to continue. If you think AI can replace your current job, testing or otherwise, it's probably time for you to start thinking about transitioning to the more human side of your profession.

Hope this helps.

Patricia Osorio Aristizábal

Dedicando más tiempo para leer más, para escribir más. A mi ritmo y con mi propia voz.

6mo

Tariq King I love your article. I agree with you on "Because testers bring a human perspective to the process and understand users' experiences." I'm trying to get back to basics: more conversations with business people and less checklists. My question is always: How can I help? (from my skills and knowledge). Thank you.

Like
Reply
Jon Bach

Award-winning speaker (2024): Enterprise Program Management | Quality Engineering & Leadership | Customer Centricity | Rapid Software Testing practitioner and advocate

6mo

I liked your article, Tariq. I respect the way you approached some key concerns a lot of us have as well. Take a look at this peer conference me and my brother are hosting and consider applying to be a presenter if you have a specific hands-on experience report to share. We want stories about how you and others brought a critical-thinking mindset to either AI-in-Testing or testing AI. https://go.sniply.page/render-assets/?url=https://www.linkedin.com/posts/james-bach-6188a811_wait-2-call-for-participation-activity-7183888973679767552-VXLv/

Like
Reply
James Walker

CEO at Curiosity Software | Driving Productivity and Quality in Software Delivery with the Outer Loop Platform 🚀

6mo

Great article Tariq. I agree, the goal should be to augment / accelerate rather than replace. Would I like Devlin to come in and fix my code / work on my boring defects? Yes. That would then free up my time to be spent on where I can add the most value - innovative solutions to hard problems, which AI will struggle with. Everyone gets so excited about AI needing to solve the hardest problems, but there is so much low hanging fruit to be had.

Jason Arbon

🧙♂️ CEO, Checkie.AI | test.ai, Google, Microsoft, Chrome, Search, “Automating the World"

6mo

Sounds like you started with the conclusion…we need to lime and hash this out

Heemeng (Chris) Foo

Leadership in Quality Engineering, Test and Engineering Excellence. Startup advisor.

6mo

Because software development is an intellectual activity where multiple people are collaborating on developing a software system (which is just going to get more complex). As with any human activity, it requires a lot of communication - written, verbal and non-verbal. Maybe when we get to the level of R2D2 and C3PO we can get there but the current state of AI is nowhere near that.

To view or add a comment, sign in

Insights from the community

Others also viewed

Explore topics