AI post-Covid-19: Have the goalposts shifted?
• Artificial intelligence (AI) is estimated to add around US$ 13 trillion to global GDP by 2030, a growth of 16% over 2018.
• The usage of AI is expected to accelerate amidst the COVID-19 crisis, particularly due to the emphasis on social distancing.
• AI can help replace humans in a range of monotonous tasks across industries, and robots can be quickly and effectively scaled up across locations.
• In these uncertain times, AI can also enable companies to make sense of both structured and unstructured data to assess emerging demand patterns and provide effective solutions in a customised pattern.
Artificial intelligence has been labelled as a game changer in the global economy much before we were introduced to Covid-19. A report by McKinsey in 2018 had predicted that around 70% of companies would have adopted one or the other form of AI activity by 2030. AI has the potential to add US$ 13 trillion to global GDP by 2030, around 16% higher than the GDP in 2018. The report added that AI can bring in major changes in the way humans live and work across the five areas – computer vision, natural language, virtual assistants, robotic process automation, and advanced machine learning.
With the spectre of Covid-19, however, something unprecedented has happened. In particular, the emphasis on ‘social distancing’ as the only true antidote implies an entirely different set of ‘discomfiting’ implications of keeping human beings at the workplace in the short term. In a bid for survival, it is only logical for companies to be looking at ways in which people can be facilitated for remote working.
However, there are several practical limitations to consider, especially in the manufacturing sectors, where companies will be compelled to work at a fraction of their capacity to keep people distant from each other. Here, there is a strong anticipation that AI could be deployed at a rapid pace by companies to stay ahead of competition. This can be true especially for tasks that are ‘monotonous’ for humans. When fed with the right algorithm, AI can take their place for these tasks. They can be quickly deployed across locations, and perform these tasks with efficiency, accuracy and 24×7.
But that is not all. As a report by Boston Consulting Group opines, AI can also help companies simulate work environments and deploy on-demand labour forces. They can detect customer patterns and deliver ‘hyper-personalised’ products. This is even more strategically relevant in the current context, where the importance of accurate demand estimation can hardly be overemphasised when it is a ‘commodity in short supply’. And this is just the tip if the iceberg when it comes to what machines can achieve, when armed with the right ‘human’ intelligence. Below, we give a brief overview of five key sectors where AI can undergo a paradigm shift in the post-COVID-19 era, and how it may transform the way these sectors operate.
As discussed earlier, robots are expected to play an increasing role in manufacturing units, as they can perform monotonous tasks without getting tired, making errors and even getting ‘infected’. Also, given the rising protectionism and disruptions in global supply chains, companies have been compelled to change their approach.
Rishu Sharma, Principal Analyst, IDC India, comments:
“Technologies like AI-powered tools can be leveraged when it comes to smooth business operations on the floor. Intelligent automation also enables monitoring machines remotely and reducing manual efforts across processes in a manufacturing setup.”
Earlier, companies emphasised on small numbers of high volume factories in low cost manufacturing destinations. Now the emphasis is on building a large number of smaller facilities that are closer to the customer. AI can play a huge role in helping companies optimise operations of such facilities, making their operations resilient and cost effective. They can realistically operate on leaner inventories, shorter downtimes and higher production speeds.
Connected factories can lead to greater integration between design, production, line and quality control for useful insights and timely interventions. Also, robotics and 3D printing can be efficiently deployed across these units to boost production when needed. AI can enable efficient management of the supply chain as well, through real-time tracking of vehicles, data-driven approach to inventory and optimisation of shipping and delivery times. A potential use case for AI is mass customisation, imparting the ability to specifically customise products as per customer needs on a large scale. Another area where AI can be of immense value is dangerous environments where deploying humans is deemed to be risky.
Just like healthcare and food, banks cannot afford to miss a trick amidst the Covid-19 pandemic. Smooth and efficient operations of financial services are extremely important to ensure continuity in business as well as day-to-day living. Critical tasks for banks at present include regular stakeholder communication, proactive assistance for account holders, compliance with regulatory requirements and adjustment to new work-from-home norms.
But banks can also see this as an opportunity towards greater technology integration in their business models. They typically have a diverse set of customers, from the senior citizens who prefer personal interactions at bank branches to a small business owner looking for growth capital to young professionals who are perfectly at ease with exclusively online banking. As they build new age digital infrastructure, AI and data analytics can help banks garner insights from their data assets to develop personalised customer solutions. This will be a core competence area in the future, which banks cannot afford to miss out.
Banks have already started using AI-based tools to replace traditional call centres and meet growing customer interface requirements. Human interface can be limited to those singular non-routine queries that cannot be handled by technology.
Dr Sriparna Saha, IIT Patna, comments on the rising role of AI in banks:
“Personalized banking can be implemented easily with the help of chat-bots, this can in turn help in improving customer satisfaction and engagement. Frequent banking queries like account balance, bank statements, transfer funds, creating a deposit, saving and investment advice, and so on can be answered by the chat-bot.”
Automation of routine tasks enables better accuracy and frees up human resources for more critical tasks. Moreover, it has helped reduce costs by 50-70% for such tasks, according to Ernst & Young.
Another important use case for AI in banking is in better estimating risks in both individual transactions and the larger financial landscape of a market. In an era where young and lean fintech companies are upending the lending landscape for instance, AI can help banks ascertain the creditworthiness of a potential borrower at much lower time and costs, and with much better rigour based on comprehensive data analysis. Similarly, AI-based fraud detection systems are showing promise, as they can analyse client behaviour and spending patterns through his/her transactions and trigger a security mechanism when there is an aberration in these patterns. Data-driven investments (also called algorithmic trading) are also showing a rapid increase in the past few years, increasing to US$ 1 trillion in 2018. This involves using both structured (databases, spreadsheets, etc.) and unstructured (social media, news, etc.) data analysis in a fraction of time, thereby enabling faster trading.
The closing of schools has been particularly worrying for students and parents across the world. Online education has been seeing some traction in the country with ed-tech startups like Byju’s and Vedantu. But these have been supplemental to the core curriculum at best, helping students grasp concepts and also have some fun learning.
Schools in general would have not vouched too much for the power of online education till recently. But now there is practically no other recourse to reach students and try and cover whatever is possible in the absence of a traditional classroom environment. But the true challenge is personalised education – the holy grail for even the conventional system of education.
A report by Global Market Insights estimates that AI in education was a US$ 400 million market in 2017, which is projected to cross US$ 6 billion in 2024. Through the use of deep learning, machine learning and advanced analytics, AI is beginning to provide some vital answers, from experiential online learning driven by AI-powered virtual role play simulations, to software for individual tutoring and neuroscience-based learning platforms.
You can visualise a BOT, for instance, that is aligned with every child from the day he/she joins school. This BOT can analyse scores of data on the child’s individual abilities, strengths and weaknesses and make appropriate interventions in the curriculum or pedagogy for that child.
Prof Anupam Basu, Director, National Institute of Technology, Durgapur, supports this view:
“So far, in online education we are approaching the students as a community or a batch, where everyone is treated in a uniform manner, but the same size does not fit all.”
This can help resolve the challenge of low teacher-student ratios and also democratise access to quality education – across geographies and economic classes.
Teachers can utilise the concept of blended learning (a combination of online and offline approaches) to optimise outcomes for students. This crisis, where around 1.5 billion students are estimated to be out of school, provides an opportunity to explore both the possibilities and limitations of online AI-based learning. AI-based systems are also showing results in grading tests for both multiple choice and long form answers. Freeing teachers from such tasks can give them more time for more strategic changes to the curriculum as well as ways to improve teaching quality, enable holistic skill development, etc.
As crowds vanish from supermarkets and malls, there is an understandable concern in a range of consumer-facing industries like food, apparel, footwear and electronics. Even though that could imply a shot in the arm for e-commerce, even online firms will have to adapt to an entirely new set of market dynamics. Consumers are not expected to be very enthusiastic about purchasing in the coming months, especially discretionary spends. Given the impact on margins, companies are expected to play the volume game. AI can help companies detect patterns of demand in the market, to make them better prepared to service customer needs. Technavio predicts that driven by the onset of Covid-19, AI market in retail is expected to grow by US$ 14.05 billion between 2019-23.
Whenever the restrictions are lifted, the entire supply chain will need time to adjust and get back to some semblance of its pre-Covid-19 stature. Demand on the other hand is expected to come in spikes, giving them little lead time to adjust. However, with deployment of AI, companies can garner insights from data on customer demand patterns and how they are expected to beave in the coming days. This will give them time to plan their sales promotions, product mix, merchandising, etc across local and global markets. Amazon, for instance, is blocking non-essential product shipments to its warehouses in the present milieu, leaving space for just six categories – baby product; health and household (including personal-care appliances); beauty and personal care; grocery; industrial and scientific; pet supplies. Accenture has predicted that AI can improve profitability in wholesale and retail industries by 38% by 2035.
With the onset of the pandemic, applications of AI in e-commerce are bound to increase. Bots can now reach customers by the millions, generate leads and even automate the entire customer experience from awareness to purchase. Virtual personal assistants are expected to make a huge difference in how customers interact with retail firms. Spending on VPAs is projected to reach US$ 2.1 billion by 2020, growing at a CAGR of over 45% since 2015. Just as the internet over a decade back, AI can act as a leveller in the market, enabling even SMEs and startups to cut sales costs and compete with larger firms.
Even in the midst of the present crisis, AI is making its role evident in several aspects of the healthcare sector. A Canadian company named BlueDot had predicted the onset of the Covid-19 pandemic in December last year. They used machine learning and data analytics to identify a cluster of “unusual pneumonia cases” happening around a market in Wuhan, China and flagged it well before the WHO.
Considering the scenario in the healthcare system where India faces a dire shortage of doctors, AI can help free our doctors from routine workloads. Machines can perform the initial role of screening and a paramedic can handle most cases using a computer. This will leave only very singular cases for the doctor to handle. Also, as Professor Anupam Basu, IIT Kharagpur, explains,
“There could be an old doctor who is very experienced and knowledgeable, but his hand is no longer steady. He knows exactly what to do, and he can make the robot do it in an error free manner.”
Robots are also being deployed to serve food, medicines, etc and sanitise areas used by Covid-19 patients. Engineers are also developing robots for functions like taking mouth swabs, doing ultrasound scans and listening to organs so that frontline medical staff is protected from infection.
Rakesh Barik, Leader, Technology Consulting, Deloitte India, adds:
“Solutions and tools are being developed by epidemiologists and data scientists using big data, and machine learning across different areas from predicting hotspots to diagnosis to even drug synthesis.”
This includes forecasting infection numbers and who will contract the novel corona virus, so that medical infrastructure can be managed accordingly. Furthermore, hospitals using AI Systems can identify feverish symptoms with a face scan. Going further, AI is also helping doctors in some countries to predict complications like respiratory failure or sepsis in COVID-19 patients.
What’s more, AI/ML is also being deployed to help find a vaccine against COVID-19. The systems search current literature on disease, study the DNA and structure of virus and then consider the suitability of various drugs. This is where natural language processing (NLP) comes in handy. Dr Sriparna Saha, IIT Patna, affirms that techniques can be applied in building biomedical knowledge graphs, which illustrate the relationship between different biological entities (such as drugs and proteins) generated after processing several scientific articles.
AI is also being explored for a deeper analysis of Covid-19 at the molecular level, helping discover novel drugs to target the virus. Clearly the experience of this crisis is expected to usher in a new paradigm in the deployment of AI in the global healthcare ecosystem.