You are here

Rise of the machines: The coming AI/testing singularity

public://pictures/jason_arbon.jpg
Jason Arbon, CEO, Test.ai

Artificial intelligence (AI) is the next exponential technology trend, and it's knocking on your front door.

For the uninitiated, AI breaks down into two major categories of study. One branch, "artificial general intelligence," is the effort to make machines that are conscious, like humans, where machines can reflect on their own existence. The other is "narrow AI"(also known as machine learning).

Machine learning focuses on computer algorithms that can be trained with data and that mimic human thinking—without actually thinking. Machine learning is rapidly growing more powerful. It can mimic the actions and judgments of humans and is already better than humans in many fields, including medicine, law, transportation, manufacturing, weather forecasting, mechanical design, and monitoring.

Even before the machines are conscious, they will be able to mimic human testers in the near future with machine learning. Given all that, what will be the impact of AI on testing? Here's what your testing team needs to know about the rise of the machines.

[ Is it time to rethink your release management strategy? Learn why Adaptive Release Governance is essential to DevOps success (Gartner). ]

The coming singularity

The AI testing singularity will occur the moment that software can test itself without human intervention. This moment will come, and we all have questions before that happens: How will machines learn to test better than humans? Which testing activities will succumb to AI first? Which will be last?

Some people say that AI is just slightly smarter software and will be unable to ever behave like a human tester. They say AI can never replace humans in testing because machines can't think, and they require human judgment and cognition to operate properly. Let me politely say that it is far easier to be a skeptic than it is to dive into the technology and understand its implications.

There will soon be a moment when software can test itself without all of us carbon-based life forms.

Some rather vocal testers say that testing requires creativity, intuition, and empathy, and that a machine can't replicate that. Despite the fields where AI is quickly eliminating highly skilled and intellectual jobs—including in transportation, medicine and law—some software testers believe that our profession is somehow immune. We want to believe that.

[ Special Coverage: STAREAST Conference 2019 ]

Key drivers toward singularity

But several factors point to the coming of an AI testing singularity. These include:

The rapid evolution of the technology

Unlike previous technological revolutions (desktop, web, cloud, mobile), AI is designed to learn from humans and to replicate their judgment and actions. Human testers make judgments and actions all day. AI focuses on solving human input/output functions. Testing is just one of them.

Heavy investments in AI

The world is spending billions of dollars on AI research, and the number of papers published on the topic is growing exponentially. For the longest time, the testing field depended on the R&D budget of open-source "charity" projects such as Selenium and Appium, and on low-priority testing efforts of tooling companies and vendors with budgets in the double-digit millions, at best.

As the testing profession builds AI-based testing approaches, we will ride this unfathomable wave of new R&D.

The fact that AI abstracts concepts, and feeds on large amounts of data and/or compute resources

AI is itself just data and compute, so AI can learn from the very data it generates in a reinforcing feedback loop, and executing this loop can only takes minutes to improve itself. All this means AI will get smarter at a far faster rate than humans can.

[ Also see Paul Merrill: Will AI bots steal your QA testing job? ]

[ Get Report: The Top 20 Continuous Application Performance Management Companies ]

A slow process

The AI testing singularity will not happen overnight. AI-based testing systems will progress slowly in the near term, through the usual cycles of hype, disappointment, then ultimately ubiquity, as all emerging technologies do.

In surveys of top AI researchers, almost all say they think AI will be on par with a human mind within 100 years at the outside. Very few think it will never happen, and many think the time to get there could be a short as 20 or 30 years.

Given that, it seems absurd that someone who isn’t well versed in AI technology, neuroscience, or behavioral science could confidently say that AI will never replace carbon-based software testing.

It is inevitable. The questions are just, How soon? and, How will it unfold?

[ Also see: 5 ways AI will change software testing ]

During my keynote at STAREAST 2019, I'll share a tentative timeline and order of when different aspects of testing progress through the three stages of "AI doesn't help," "AI helps," and, ultimately, "AI does it," across the varying aspects of the software testing profession. I'll explore this roadmap so you know how and when to bring AI into your own testing efforts, and know where to invest in your personal learning to stay ahead of the technology curve. TechBeacon readers can save $200 on registration fees by using promo code SECM.

[ Get Report: Buyer’s Guide to Software Test Automation Tools ]