Dr. Xin Xu – Award-winning Expert and Leader in AI, ML, Data-driven Decisioning – explores the growth in AI, ML and AR during a crisis pandemic – from healthcare research to call centres and IT help desks.
Major disruptive events, like natural disasters, are often the catalyst for great innovations and moving emerging technologies into mainstream use. Such is the case with Artificial Intelligence (AI) and its subsets of Machine Learning (ML) and automation in the COVID-19 pandemic.
AI is playing a direct role in the medical and healthcare domains, as well as in many of the ways the pandemic has impacted our everyday work and life – from new remote or virtual work environments, to pressure to “do more with less” and physical social distancing requirements. Is AI truly a hidden hero finally revealed, or is it merely a fad? Here is my view…
The COVID-19 crisis has highlighted many practical uses of AI in the healthcare sector, including ML-led drug discovery, AI-powered epidemiology systems that detect outbreak and issue warnings, autonomous robots delivering food and medication to quarantined patients, AI systems that support diagnosis and clinical characterisation. In fact, scientists have applied AI/ML to medical researches and bioinformatics for decades. AI/ML is now an indispensable tool and standard practice in studies of virus, diseases, genes, molecules, drugs and related areas. This trend will continue, and will likely be intensified by the COVID-19 outbreak. Already MIT Technology Review’s “10 breakthrough technologies in 2020” predicts AI-discovered drug molecules as one of the breakthroughs to be expected in the near future.
In the midst of the pandemic, when the demand for information, analysis and assistance is at its peak, AI helps call centres to complete more workloads with less resource, typically through “Intelligent Process Automation” (IPA). While IPA has existed for some time, the current pandemic has expedited its progress to improve and operationalise IPA in the real world. For example, the virtual agents (aka “bots”) used in contact centres or service desks for everything from government services to financial or IT service providers, now understand users’ requests better to take smarter, faster actions. In unusually high peak periods, these bots significantly help by taking the burden off human operators to reduce call wait-times and improve the customer experience. In normal days, it also helps optimise routine processes. One example of IPA in the area of omni-channel digital Workplace Services, is in AI-Ops platforms such as Unisys InteliServe™. They use chatbots in conjunction with IPA to resolve or handle IT issues in a self-service model, without the need for users to wait for or interact with human operators. In this way, AI is changing the operating models of IT service desks and contributing to the welfare of both workers and users.
Another pandemic example is the demand for remote or virtual support, driven by behavioral changes like social distancing requirements. As a result, Augmented Reality (AR) is emerging as a way to provide support, such as enabling IT technicians to provide support without having to travel to or enter the physical locations, using smart glasses or via an app on a mobile phone instead. When combined with AR, AI may provide better recommendations or real-time alerts during the remote work, by leveraging the big volumes of data collected by AR devices, much like interactions between AI and IoT devices.
At the core of all the different forms of AI, is the AI/ML-led “autonomous decision engines”. What I mean by “autonomous” is that the engines will create their own rules about what actions to take, rather than adopting pre-determined rules set by human experts. These engines can conduct specialist tasks and achieve relevant KPIs comparable to, or exceeding, the level of human experts, when they use AI/ML trained by past data to design the rules incorporating deep domain knowledge. More importantly, using combinations of different, specialised decision engines (listen to this podcast), key processes or operations may be intelligently automated, rather than mechanically automated. It can run 24/7 in absence of human rules; detect anomalies or infer patterns never seen before; and continuously improve itself; within a governance structure that potentially gives the control back if human interventions are necessary.
Today it is common to embed AI/ML-led intelligent decisioning endpoints within automated operations or robotic processes. This “embedded AI” means many people may not even realize that AI/ML is the backbone of their daily decision-making. However these silent efforts should not be underplayed– especially during the pandemic. It helps meet users’ pragmatic and result-oriented requirements, and addresses their primary concerns of accuracy, security, timeliness, productivity.
The COVID-19 pandemic is having broad and profound impacts on whole world, some of which may be far-reaching even long after this event. AI and ML emerges as an important tool to offer helps on many aspects of our life and work. This in turn provides a good opportunity for AI/ML to grow. I think AI is a sustainable and promising trend, for now and in the post-pandemic era. On the other hand, we shall not take booming of AI for granted – we need to consistently innovate and operationalize AI in our daily business and work, in order to reap its benefits ultimately.