‘Arti-fictional’ testing? – The journey from CAPTCHA to Skynet through Turing Test.
Many multitudinous mundane tasks are getting slowly but steadily augmented by Artificial Intelligence. These repetitive tasks are easy to replicate through machines and smart programming because they are easy to model through code.
In many avenues of the world, testing is a repetitive and easily programmable task that a lot of people are working towards perfecting for their reasons and aims. Artificial Intelligence based testing can execute calculations that are profoundly more complex and comprehend far more data and attributes that a human tester can. Computers can consume datasets better and faster than Joey Tribbiani consumes food.
Computers that are programmed well and trained on the right datasets can outperform humans by a large margin, one that is impossible to ignore. Many teams around the world have observed multi-fold increase in productivity and testing capabilities.
The Chinese Room Argument:
There are many kinds of tasks that humans still excel at over computers, one such task is thinking. It is because of this thinking prowess that we are able to program such machines. Moreover, it is this thinking process that the programs that we think of aim to test.
That is why we make computers through thought to test another’s thought capabilities. However, it does not quite equal that of a human testing a human. At least that is the premise of the Chinese Room Argument (pun intended?).
To sum it up concisely, the Chinese Room Argument believes that a weak/narrow AI can only ever process the syntaxes and can never truly understand the semantics of the language. It will always be able to stimulate thought and to an unobservant mind, can maybe even give the illusion of free thinking.
We have come to the point that is a reasonable distance from that of a weak AI. We have now seen computers come close if not already break the Turing Test and we are witnessing a spectacle in terms of computer intelligence with Google, Amazon, Apple and such companies leading the way with their skillful AI software.
You may like to watch a Video on Ai in Test Automation by Indium Software
The New Way:
Artificial Intelligence based testing has brought Automated Software Testing from simple Q/A systems to more complex systems and now to near sentient testing arenas where the feel of the test simulates that of being examined for a viva. Many more subtly profound tests use cognitive computing and also beneficiary for Software Testing Services Companies.
For example, the GMAT where following questions get harder but also give you higher scores depending on how you solve the preceding questions.
There are many hints and tinges of cognitive computing that has slowly but steadily changes testing processes all over the world. We don’t even realize many of these changes that happen. But they are happening, and they will continue to get better and more prevalent.
I came across a different sort of AI-based testing through a start-up. They were an EdTech startup called Leverage. Leverage used Machine Learning to match-up each student with a mentor for further personalized learning and guidance. That does not seem like testing, but it is, in a certain way. One’s attributes are tested and matched, and then a satisfactory conclusion arrives.
The list of examples keeps going on, but you get the point. Artificial Intelligence, Machine/Deep Learning, and Neural Networks have all coalesced to create a brilliant future for cognitive computing where we can soon tangibly realize a Jarvis or a funny Deep Thought and possibly avoid a Skynet or Ultron.
Author: Jessica Cyrus