By Kalilur Rahman
“AI (Artificial Intelligence) is the new electricity“—Andrew Ng (AI Guru, Founder Coursera, Ex- Baidu).
Ever since the term Artificial Intelligence was coined by John McCarthy in 1956, many innovations have taken place. With the increasing innovations in the landscape over last 60 years and explosive growth in the last decade, there is a varied level of understanding of the concept. Alan Turing, one of the greatest computer scientists made an accurate prediction regarding the development of AI. Humans have grown leaps and bounds since his initial prediction.
Nvidia offers a comprehensive diagram that depicts the growth of AI in 3 distinct parts—AI, Machine Learning and Deep Learning and how it has progressed in the past 60 years. AI has been able to beat the best players in chess, ‘Jeopardy!’ (A Top US Quiz Game) and ’Go’, one of the most complex strategy game. However, they failed in some cases which showed that they can miscalculate certain basic moves or questions which resulted in a human victory.
The growth possibilities are limitless
The real question, however, is the extent to which AI can grow. The technology is feasible and the possibilities are limitless. Certain AI Systems can write algorithms and programs to reduce the development time. They can also invent a new language for processing like the Facebook bots. There are trainable AI bots that could be as fragile as humans with biases (like the Microsoft Twitter bot that got shut down after a racial cognition error).
Many of these tools have been in existence at a much lower scale and capacity such as the Integrated Development Environments (IDEs) that help in writing faster codes, CASE (Computer Aided Software Engineering) Tools, code generators, etc. However, the increasing intelligence and potential of these tools will take us to the next level. Recently, a YouTube and a WhatsApp video of a Japanese humanoid robot crossing the street with typical human emotions and moves created a big stir and went viral. Hence, practical use of these technologies is real and useful in a variety of ways.
There may be a digital disparity when it comes to adoption of AI. However, AI Startups are growing and have become a hotbed for venture capitalists in industries delivering AI driven bots, automobile, computer vision, core or functional AI, commerce, Internet of Things (IoT) and Industrial Internet of Things (IIoT), healthcare, financial technology, robotics, analytics, cybersecurity, and sales and marketing.
Testing substantially improved through AI
Various statistics have shown that faster testing mechanisms have become the need of the hour. The world is moving away from yearly, quarterly or monthly testing process to minutes and hours based testing mode as explained by Tricentis. ‘Intelligent Machines’ or ‘Artificial Intelligence’ is used extensively by all the top technology firms to propel their dominance to the next level in their industries. AI will help in many ways to reduce the pain of ever-increasing coverage gap arising due to the two key drivers—increasing complexity and reduction in time available to test. AI can also help in addressing the key questions that testing leaders face regarding the balance between time, scope, cost, and quality and the resources, test coverage, test efforts, and test costs.
Concepts like reinforcement learning, passive reinforcement, regression algorithms, linear regression and much more can be applied for leveraging all these techniques. Out of these, Monte Carlo tree search, support vector machines, search trees or deep neural networks, Bayesian classification and clustering, and probabilistic models can accomplish numerous advanced techniques. These include automated defect detection, automated exploratory testing, test coverage heatmap, self-healing automation, self-adjusting regression, pattern recognition for various metrics and trends, risk and coverage optimization, diagnostic, prescriptive and predictive analysis, and predictive modelling for test design defects among others. One of the ways an AI model can be built up for testing is by feeding inputs for learning and enhancement of the models, building algorithms for Deep Learning and enhance the engine on an ongoing basis.
King’s AI usage in testing
King Digital Entertainment plc, doing business as King, is a social game development company. The company offers an interesting approach to AI usage in testing. They make well-planned use of deep artificial neural networks, Monte Carlo tree simulation, bots to perform testing, and AI engine for a continuous feedback loop. They have undertaken advanced automation by bots and regular updates of machine learning models.
A hybrid test team (150-200+ testers) with unique skills covering test automation and data science programming along with the use of data scientists for domain knowledge ensures regular crash testing, performance testing, regression testing as well as regular maintenance of AI bots for testing.
AI will enter mainstay of the next wave of digital revolution given the popularity of Candy Crush Saga, how King leveraged AI to do testing efficiently, code refactoring and quality engineering by behemoths such as Google, Amazon, Facebook etc.
Implementing AI, but judiciously
AI can be used to undertake exhaustive testing such as non-creative, non-value adding, effort-intensive, repeatable automation and data creation, predictive exploratory testing etc while the qualitative aspects must be left to human creativity to ensure continuous innovation by humans. Any decision must be taken after taking into consideration the relevance, usefulness and long-term business and customer centrist gains. Organisations should automate or ‘BOTomate’ early for getting returns from early phases itself.
However, there are two sides of the coin and AI is subjected to a lot of scepticism. In order to innovate, new ways have to be found to automate and propel capabilities by leveraging the opportunities given AI. At the end of the day, it will be augmented intelligence, adaptive intelligence, and automated Intelligence that will propel human intelligence forward. Whether it would be an “Eye (AI) Wash” as sceptics say or an “I wish” from them for starting late on the journey, only time will tell.
Featured Image Source: Pexels
Stay updated with all the insights.
Navigate news, 1 email day.
Subscribe to Qrius
