CIOInsider India Magazine

By Keerthana Kantharaj
 

Ordering food online, unlocking a phone with your face, or even talking to your phone saying ‘Ok Google’ or ‘Hey Siri’ and getting a human(ish) response were just dreams a couple of decades ago. But thanks to the ensuing experimentations with artificial intelligence and machine learning, among other technologies, it is not only that all these are possible, but these buzzing technologies are digitalizing the way we travel, work, communicate and live.

Today, there is high demand for individuals skilled in working with the latest technologies, and an average ML engineer earns about Rs.7.6 lakhs, according to Glassdoor’s data. Likewise, the global machine learning market is expected to reach $152.24 billion in 2028 by 2028, growing at a CAGR of ~38 percent, according to MarketResearch.com. This is primarily due to the upswing caused by the COVID-19 pandemic that shook economies and fast-tracked the need for technological solutions for a better living. Engaging in an insightful conversation with CIO Insider to decipher the latest trends in the ML industry is Navin Nathani, General Manager and Head - IT and Transformation, Aditya Birla Group, Hindalco Industries. Navin is a well-endowed digital-first enthusiast and award winning techno-functional IT business leader with 20+ years of experience leading digital transformation and integrating business and digital processes. With a stronghold of IT business experience, Navin explores the realm of machine learning and its impact on human life today.

In conversation with Navin Nathani, General Manager and Head - IT and Transformation, Aditya Birla Group, Hindalco Industries Ltd.

How do you perceive the flow of investments in AI and the current growth factors driving the ML market?
Investments into AI globally have risen by 115 percent since 2020, estimated to be the largest year-over-year growth. The total investment reached about $77 billion last year, a substantial increase compared to the previous record of $36 billion. Clearly, COVID-19 is mainly responsible for drawing global interest around AI from both governments and investors. Since the situation called for an immediate need for digital collaborative spaces and remote working tools, forcing business leaders to understand digitization. The US stands as a global leader in AI thanks to its leading commercial market, large talent pool, and stellar research initiatives. China takes the second stand for supporting the administration of the technology, including ML technologies. The third is the UK for its increasing growth in ML, primarily in the next generation of computing architecture.

Additionally, machine learning training processes are accelerated by the huge cores accompanied by graphics processing unit (GPU) systems. These systems help enhance perfect decisions for perplexing ML models. They are available widely across consumer devices to ML in the public cloud.

Likewise, cloud computing and organizations migrating to the cloud are becoming mainstream. As a result, data accessing costs are now reasonable, with organizations investing heavily in the cloud computing space. Since data scientists require access to huge sums of authentic data for preparing ML models, they are capable of anticipating increased precision. This is required to take care of complex issues such as distinguishing growth or downfall, for which analysts need substantial datasets with diverse data points. Therefore, organizations are leveraging inexhaustible data added with high-performance computing gadgets, driving ML and AI solutions.

Simultaneously, businesses began realizing the power and potential of ML and analytics. It not only helps to understand the customer but also creates agility in their supply chains and distribution networks. As a result, it specifies the need for a data foundation to oversee datasets to put that into an insights engine using algorithms and to activate new revenue channels.

Another factor driving the adoption is the access to high-quality, robust and adaptable machine learning models, rapidly designed and accessed thanks to open source frameworks like TensorFlow, Python and scikit-learn.

Also, organizations adopting ML can become more agile and customer-centric by pairing their workforce with the right skill sets to seamlessly implement AI and ML.

How equipped is the talent pool to leverage these latest technologies?
I’d say there is definitely a great room for improvement. Although India is known to house the highest number of engineers, industry data reveals otherwise. The country grappled with sufficing technology resource requirements for about three to four years, exposing a demand and supply gap in digital talent, estimated to be around $ 6 million. This is due to the abrupt upswing in demand for professionals fluent in skills such as AI, cybersecurity, ML, blockchain, digital marketing, and cloud computing, among others. The India Skills 2021 report proves that only about 45 percent of educated candidates possess the aforementioned skills.

The situation turned grave, struck by consequences of widespread great resignation. The attrition rates of IoT companies remained at an all-time high, with quarter-over-quarter attrition rates ranging between 15 to 25 percent, eventually leading to a cutthroat talent war among industries.

A sustainable solution to bridge the skills gap would be investing in upskilling. This can be done by training graduates from the science field in digital disciplines such as blockchain, AI, ML, and data science while encouraging them to take up new roles.

On the other hand, edtech companies began developing cutting-edge learning tools such as technology-led learning and interactive sessions with experienced faculty based on classroom format. Not only did this attract professionals to the job, but it also helped maintain industry competitiveness while adapting to dynamic work environments and future-proof an alternate talent pool.

What is the kind of role played by ML in the Metaverse?
The answer is that it's doing everything. Take the Location Apps, for instance, which suggests places like coffee shops or bars based on the time it's accessed. A few trips later, it starts suggesting whether you should head home or party somewhere else Our history of interactions with the world, collected and synthesized by the technology we use, fuels inferences made about us. Art, defense, realty, and myriad other industries are already bracing for the interoperable VR jungle. Video games like Roblox and Fortnite and Animal Crossing: New Horizons, in which players can build their own worlds. A Travis Scott concert held within Fortnite last April drew more than 12 million concurrent views, ML is revolutionizing every part of the economy designed for the metaverse. Proving essential to the new hyper world is the logical next step, especially since more leaders are turning towards the metaverse.

What are the barriers organizations have to cross to transcend into the metaverse?
Securing sensitive intellectual properties, financial documents, and customer information is vital for maintaining business health and competitiveness. However, cybersecurity is an incredibly difficult challenge. Each business has different information security requirements, and qualified IT professionals specializing in the field come at a high price. Meanwhile, malicious actors are constantly devising cyberthreat strategies.

Secondly, businesses have to adopt compliance with a wide variety of complex compliance requirements. It’s challenging as it depends on the company’s size, industry, business model, and customers. For example, companies serving customers from the European Union must follow GDPR (General Data Protection Regulation), PCI DSS (Payment Card Industry Data Security Standard) for businesses within customer payment solutions, and HIPPA (Health Insurance Portability and Accountability Act) requirements for healthcare businesses.

The third challenge for organizations is to steer clear ahead of the digital transformation strategy. Organizations need to assess and reassess the reason for replacing legacy systems and manual processes with digital systems. These are some of the questions organizations need to answer before rolling out a digital transformation process. It's important to prevent organizations from buying into false assumptions or buzzwords. They should know the areas that require a digital upgrade.

Fourth comes the cultural mindset, since organizations with legacy mindsets, following manual processes, often appear to have an old-school mentality. They tend to change slowly, as they perceive new technologies as difficult to adopt and negatively impact automation. Therefore, everyone from the leadership to the new employees must be on the same page, ready to take on big changes in their day-to-day lives.

The fifth challenge is the IT Budgets. Scope creep tends to delay deadlines amid the addition of consultation work, changes in customer needs, and IT errors, while the cost of digital transformation goes uphill.

This leads to another challenge: the change management strategy. Organizations with a clear change management strategy are six times more likely to take the lead or win the race towards meeting digital transformation objectives. However, a lack of change strategy could result in any new project or implementation plan backfiring.

Could you give us a picture of the future of the ML market?
According to Gartner, there will be 2.3 million jobs in the AI and ML field this year. Amid evolving use cases in the area of public governance, researchers from IIT Jodhpur and Western Michigan University developed neonatal and infant mortality predictors using machine learning techniques. The ML algorithms assess important characteristics such as being firstborn, being born in economically backward households, and having low birth weight. The indicators do not use advanced medical knowledge and are easy to use for community health workers. Providing timely medical care to reduce the infant mortality rate in India is one of the use cases showcasing ML’s optimization operations.

Given these optimizing operations, it is possible to expect the adoption of emerging ML technologies, promising to enable retail stores to monitor body temperatures through thermal imaging and computer vision. Additionally, manufacturing operations are already being optimized across supply chains through sensors and IoT technologies.

Furthermore, AI and ML are being deployed in the renewable energy industry to mitigate the unpredictability of sources. ML’s use case in healthcare involves robots performing complex surgeries, reading patient records, and devising personalized treatment plans.

In agriculture, variable rate technology (VRT) is used for plant disease prevention.

The financial sector uses ML algorithms to prevent fraudulent activities based on historical data to predict fraudulent transactions. Secondly, ML and computer vision algorithms are also used to prevent identity theft by checking for identity matches in key databases.

Moreover, retail, social media, and entertainment platforms are increasingly adopting mass personalization, similar to e-commerce and media platforms using hyper-personalization experiences.

Therefore, by 2026, we could expect to witness about 60 percent of enterprises in India combining human expertise with AI, ML, and pattern recognition, to increase worker productivity and efficiency by 20 percent. By 2025, 40 percent of India companies using computer vision will use a pretrained model in a low-code environment that either fits their need or can be adopted with transfer learning from sparse data sets. AI and ML will commence taking over routine tasks across healthcare, BFSI, and other industries while maintaining competitiveness. Likewise, we need to upgrade our skills to leverage these new ML and AI-driven workspaces, which will be the new realm of digital transformation in the future.

Disclaimer: All the views expressed in this article are that of the author and doesn’t represent the views of the organization. To help readers, information has been sourced from various internet articles and credit for the same goes to respective publishing organizations. The magazine publishing this article has also sourced information from internet and credit goes to respective publishing organizations.