Is Your AI Intelligent Enough for a Pandemic

A lot has been written on how Digitisation and AI will see a surge due to the Corona Pandemic and that data driven decisions can mitigate the challenges posed by the pandemic and therefore make organisations more resilient to the current crisis and future ones. However, whilethere is merit in this thinking, we should also pause for a moment to consider the limits of the ML model when confronted with unplanned events like the Pandemic and the subsequent economic crises.

The current pandemic has adversely affected artificial intelligence, causing hiccups for the algorithms that run behind the scenes in inventory management, fraud detection, marketing, and several other fields.Machine-learning models trained on normal human behaviour are now finding that the“normal” has changed, and several of them are no longer working as they should. 

To better understand why unusual events, challenge AI algorithms, we need to understand that most machine learning algorithms, including deep neural networks, basically share the same core concept: a mapping of features to an existing model which is then used to predict outcomes.

For example, if we take the case of an automotive supplier who is integrated in a global supply chain. When embarking on advanced analytics solutions they would have begun with a simple regression on some time series. To better manage the various global factories the next step might have been to create a machine learning algorithm that predicts the sales of each unit. With enough data, the machine learning algorithm will then create a formula that can predict sales and improve the inventory management of finished and unfinished parts.

With the passage of time, the company would probably become dissatisfied with the predictions of a very simple machine learning model, since they aresometimes not very accurate.Therefore, to improve the model’s performance, with some feature engineering the company will add more variables to the model like regional peculiarities. While retraining the machine learning model, it is likely that patterns will emerge, that will make the new AI algorithm much more flexible and resilient to change. It will also be able tomake more accurate predictions than the simple machine learning model that was limited to a basic time series.

However, there is still one crucial aspect which artificial intelligence algorithms do not take into consideration.  They do not understand Causation. Their function is limited to revealing corelations between variables. This doesn’t pose a problem as long as nothing hugely unusual happens. Singular irregularincidents may disrupt the situation temporarily; however, it soon bounces back.It is a black swan event like the current pandemic which ishugelyunusual and has major and unpredictable impacts across all sectors and geographies that throws the entire system out of gear. The narrow AI systems, that we have today, are not very good at dealing with the unpredictable and unusual. However, AI is not the only system that is failing. Fraud detection systems, spam and content moderation systems, inventory management, automated trading, and all machine learning models that had been trained on our usual life patterns are breaking up.

While we humans are also puzzled and confused when faced with unusual event, we have however been blessed with intelligence that extends way beyond patternrecognition and rulematching. We have all sorts of cognitive abilities that enables us to re-invent and adapt ourselves to our ever-changing world.

These cognitive abilities are what come into play during a crisis, and the compassion evinced, distinguishes man from machine. If we take again the automotive industry as an example, as a matter of fact the Corona Virus Pandemic has resulted in several automotive factories being shut all over the world. However, several OEMs like GM, Hyundai, and Ford were volunteering to use their factories to manufacture several products needed by their countries. With their command of legions of supplier companies making plastic and metal parts as well as electronic components, it was possible to set up on short notice manufacturing processes to produce all kinds of immediately needed material from simple masks to ventilators.

The connected world we are living in has resulted that the impact of the pandemic could be felt far and wide, touching mechanisms that in normal times remain hidden. If we are looking for a silver lining, then now is a time to take stock of those newly exposed system defects and ask how they might be designed better and made more resilient. It is clear that if machines are to be trusted, we need to watch over them. However, more and more businesses are buying machine-learning systems but lack the in-house expertise needed to maintain them

Final Thoughts

Whatthe pandemic has clearly revealed is how intertwined our lives are with AI. There is a delicate co-dependencywhereby, changes to our behaviour change how AI works, but at the same time AI will also impact our behavioural patterns. This is also a reminder that human involvement in automated systems remains key- retraining a model requires expert human intervention.

For the present, what we have are AI systems that can perform specific tasks in limited environments. One day, maybe, we will achieve artificial general intelligence (AGI), computer software that has the general problem-solving capabilities of the human mind. That’s the kind of AI that can innovate and quickly find solutions to pandemics and other black swan events.

Until then, as the coronavirus pandemic has shown us, artificial intelligence will be about machines complementing human efforts, not replacing them.

About the Author

Markus Pfefferer, Managing Partner at Tibil Solutions has particular expertise in global strategic planning throughout several industries including technology, renewable energy, automotive, heavy engineering. Specializing in market development and implementation of business plans, product development, and management of international business development; Markus Pfefferer has been influential when partnering with clients in establishing business strategies throughout Europe, Middle East and Asia Pacific.

About Tibil Solutions

Tibil Solutions is a data engineering and analytics company enabling organizations leverage data for competitive advantage. Tibil enables businesses from Banking & Financial Services, Retail, Healthcare, Manufacturing, Social Sector verticals to achieve better agility, growth and profitability. Founded in 2010, TIBIL is headquartered in Bangalore with offices in Illinois and Hyderabad, India. Tibil serves more than 60 clients in USA, Middle East, APAC and India and has a strong leadership team with collective experience of 150+ years and pedigree of working in F500 companies.

Read More News :

Cloud4C collaborates with Google Cloud to build Centre of Excellence

How to keep your tech business relevant in an increasingly competitive industry