Artificial Intelligence (AI) has become a technological reality for businesses and organisations across the globe. AI can boost process efficiency, minimise human errors and workload, and extract insights from Big Data, making it one of the most widely discussed technologies today. In India, various sectors are adopting AI technologies and taking innovative steps in their applications. India is one of the world’s fastest-growing economies and has a considerable stake in the AI revolution. According to Accenture’s latest AI research paper, AI is predicted to boost India’s yearly growth rate by 1.3 percentage points by 2035.
The Indian manufacturing sector has started adopting Industry 4.0 practices, where AI can play a critical role. The pace of adoption is only expected to rise as varied sectors begin to digitise and automate the processes in the post-COVID-19 world.
AI in the manufacturing industry
AI-based solutions are projected to have a significant impact on the manufacturing industry. As per industry research, the global AI in the manufacturing market size was $840 million in 2019. It is expected to reach $4,798 million by the end of 2026, with a CAGR of 28.1% during the forecast period 2021-2026.
As mentioned earlier, the adoption of Industry 4.0 practices is a key reason driving this growth. AI is a critical component of Industry 4.0, as it enables flexible & adaptable systems, automates manufacturing & business processes and helps in making intelligent decisions. Manufacturing organisations can immensely benefit from using AI in engineering, R&D, supply chain management, production, maintenance and in-plant logistics & storage. AI technologies have significantly advanced in the past decade on several factors. It is more democratised than ever before, is much easier to understand & implement and it provides better & consistent results, thanks to the improved computing nodes, availability of trained & skilled resources and improved AI ecosystem. All these have contributed to the rise in AI adoption in Indian manufacturing industries.
The technology can also enhance quality inspection and accuracy. For example, AI algorithms can significantly improve the facial recognition of quality inspectors, thus, ensuring the correct human inspector is auditing the quality of the products. Humanoid robots, enabled by AI/ML algorithms can take over the inspection of the products and will deliver a high quality & consistent throughput.
Application of AI in cobots
Another use of AI is in cobots – collaborative robots that can work alongside humans and take care of repetitive tasks. They need AI/ML to evaluate data to create a vision for themselves and not hinder humans while collaborating. Further, AI/ML helps cobots work better in a collaborative environment continuously.
Much like cobots, AI/ML plays a crucial role in the working of Automated Guided Vehicles (AGVs) in the indoor environment, such as manufacturing factories or huge warehouses. AGVs help carry heavy loads on the shop floor. Through AI/ML, they observe the surroundings, process the captured data and move seamlessly without creating any obstacles to the humans & other equipment. Today’s manufacturing plants collect massive quantities of data, aided by connected devices and smart sensors. However, AI technologies are required to make sense of this data to generate meaningful information on the plant operations, the health of the equipment and other business insights. For example, AI helps detect issues with machines in advance before it breaks down, thereby saving the downtime cost. This is predictive maintenance and is one of the fastest-growing applications of AI in the manufacturing industry.
By using AI in analytics, companies can quickly figure out the gaps and improve their processes. Areas of supply chain management, production, in-plant logistics and storage, etc, can hugely benefit from AI-powered analytics. Chatbots are another standard tool where AI is used. They can resolve day-to-day problems on the shop floor and become an interactive way to access resources when needed. Apart from these, manufacturing companies are also using AI in their engineering and development. For instance, to develop simulations and test cases for specific functions.
However, some critical questions that arise are, is AI completely reliable? How much automation can AI do? Can AI understand certain scenarios and take appropriate decisions as humans do? And many more such questions that need understanding.
The potential of human-centric AI across industries
As demonstrated by the use-cases mentioned above, AI is all about moving from programming and more towards Machine Learning (ML). AI technologies are trained to learn various processes and patterns of a specific function and decide how to proceed to the following stages through Machine Learning. The quality of AI programs is directly proportional to the amount of data that it is trained on. To provide accurate results, AI programs need a very large amount of data to learn than humans.
However, AI can create a slew of problems. AI programs are designed to assess and overcome a specific set of loss functions, but this has unintended consequences for other loss functions. AI also has a range of unforeseen effects in terms of ethics, biases and privacy, all of which necessitate close analysis and goal-directed action. If the initial dataset fed to the system is biased in any manner, the machine will learn to generate biased outputs.
Taking a slightly non-manufacturing example, if only men were selected for manufacturing roles earlier, AI is bound to choose more male CVs in the new positions too. It might not understand that earlier selection was due to a lack of female candidates and not gender preference.
Human-centric AI is a new concept being explored to counter these side-effects of AI. In simple terms, human-centred AI refers to systems that improve over time because of human input. It applies to manufacturing and almost every other industry where AI plays a crucial role in day-to-day operations. It extends the boundaries of previously limited Artificial Intelligence solutions to bridge the gap between machines and humans by developing machine intelligence to understand human languages, emotions and behaviour and focuses on algorithms within a more comprehensive, human-based system, learning through human input & collaboration. It refers to systems that improve over time due to human input, while also offering a positive human-robot interaction.
Human-centric AI has enormous potential across industries, including automotive and manufacturing. In 2020, India joined the Global Partnership on Artificial Intelligence (GPAI) as a founding member to support responsible and human-centric development and use of Artificial Intelligence.
Human-centric AI is the future of Artificial Intelligence
To get optimal results, many procedures rely on more than just simple data. A good AI solution must be able to comprehend human desires. Specific human characteristics that are typically required for decision-making are absent from AI systems. For instance, it lacks empathy and a basic understanding of social & cultural issues. It may be missing the context required for optimum results. With consumer-facing AI systems, this becomes even more challenging.
Humans tend to make mistakes. However, AI systems can also make mistakes if it doesn’t have all the knowledge it needs to make good decisions. In the manufacturing industry, an AI system’s dependability and efficiency may be harmed by the lack of human-centric design.
Human-centric AI integrates AI’s analytical capability with humans’ creative problem-solving and expertise, thus successfully bringing the best of both worlds together. To avoid feedback loops and erroneous outcomes, consumer-facing AI systems must take a human-centric approach. Human-centric AI also improves industrial business.
To summarise, human-centric AI is essential for smart manufacturing as it increases reliability, improves decision-making and output efficiency.