Artificial Intelligence (AI) is a powerful technology. And so far, we have only scratched the surface of what may be possible as we continue to develop AI solutions. It has an important role to play in helping manufacturers to prevent accidents, run smarter and significantly enhance the safety of operations.
Using AI technologies for safety is vital for organisations in industries, like the manufacturing industry, energy, oil & gas, chemicals, mining and minerals that can run into hazardous situations. Industrial accidents & incidents associated with disabilities & fatalities in India are major concerns in manufacturing sectors. The major causes of accidents come about during the operation of machineries, industrial boiler explosions, fire & harmful gas leakages, etc. By using data to help identify patterns and determine the root cause of an accident, incident or near miss, AI systems in manufacturing are aiming to prevent accidents by giving workers the tools to predict and thus, prevent accidents, and most importantly, fatalities.
AI & safety in manufacturing
Rajesh Ramachandran, Chief Digital Officer – Process Automation, ABB, says that in India, there is significant room for improvement in various aspects of safety. Analysts suggest that 40% of safety incidents can be attributed to injuries due to falling which can be easily prevented. “By combining IoT technologies, real-time data collection and with advanced analytics, AI helps to improve the safety of personnel & physical assets within a manufacturing environment,” he explains and goes on, “AI, in the context of safety, covers a wide spectrum, from collaborative robots (cobots) to machine vision technologies as well as algorithms & models for Machine Learning (ML) & deep learning. These technologies are highly promising in terms of the benefits they can bring for safer operations through a strategy of ‘integrated safety excellence’.”
Thenkurussi ‘Kesh’ Kesavadas, Co-founder, AirV Labs and Professor of Industrial Engineering, University of Illinois, USA, goes on to mention that one of the very common causes of accidents in the industry is non-compliance, like people not wearing harness, working in unsafe areas or at heights, etc. So, non-compliance is something that can be monitored through AI systems. “For example, a computer vision system or sensors can monitor how people perform inside the factory. Early indications can be provided if people are not following the regulations, which can send warnings or emails to managers and cut down improper practices, thereby improving the safety situation inside a factory,” Kesavadas expounds.
He further goes on to assert that another practical way of improving safety is through wearable sensors, which can be worn over a jacket, on the wrist or on the helmet. These sensors can be streaming data & monitoring activities that are going on in a situation in the factory that might be hazardous. If one sees occurrences through Machine Learning and data analytics that are not routine, that can also create a warning before a serious accident can happen. Just like in automated vehicles or self-driven cars, these kinds of technologies can also be applied inside a factory where multiple forklift machines can actually talk to each other through sensors and 5G technologies. They can avoid collisions and hitting objects inside the plant.
“Fundamentally, the way AI should be used is more predictive,” Mahesh Subramanian, Co-founder & CEO, CamCom, puts across and adds, “And as long as AI starts becoming more predictive, then one should go back and start fixing issues at the core rather than reacting to issues later in the supply chain. That is how the correct use of AI would work in the manufacturing segment. AI is never going to be 100%. So, the prime focus should be to get to 90-95% with AI, while at the same time, training humans to avoid the other 5-10%. Therefore, AI can eliminate most of the repetitive tasks but if there are any higher order tasks that require a kind of a thought process, then AI is still not there in this point in time to do that.”
So then, while humans are trained, what role do they play when it comes to AI? Moreover, we ask how they can join forces with AI to boost each other's strengths and bring accidents to the minimum in the manufacturing industry.
Humans and AI
“If any task is highly repetitive and monotonous, then those are the tasks best suited for an AI system to take over and the higher order thinking skills that are needed are brought in by the humans. Because humans get bored easily,” Subramanian further elucidates and continues, “So, if we are asked to do the same thing over and over again on an eight-hour shift, then we get bored and fatigue starts to set in. And then we start making mistakes as humans, which is where machines have to work in conjunction with us.”
Being on the same wavelength as Subramanian, Suresha Parashivamurthy, Manager – Artificial Intelligence & Machine Learning, Faurecia India, tells us that AI technology and robots don’t get tired and neither do they get distracted, slow their pace or get bored. “This doesn’t remove humans from the forefront of work, though,” he asserts and goes on, “Instead, humans continue to play a role in supervising, programming and repairing robots that are performing repetitive or dangerous tasks. And though it might seem that using AI technology to prevent accidents & protect workers will lead to a human-free job site, it’s doubtful that that will ever be the case. Robots will still need to be supervised, and human workers will find that they still need to ‘teach’ robots in some cases, to properly complete a task.”
Are 100% zero accidents really possible?
With AI’s capabilities and the humans’ bit in enhancing it further (and vice-versa), AI has a prominent contribution to increasing safety in the manufacturing industry. Although, many are of the opinion that it is not possible to truly achieve 100% zero accidents in manufacturing through AI. For instance, Kesavadas believes so because, very often, the human beings are the weak link in any AI or machine control systems. So, one can develop systems that can ideally solve most of the problems but if people don’t follow the rules, then AI cannot actually do much. Similarly, Ramachandran thinks that it is not reasonable to expect that any amount of employee training, AI technology or protocols can provide a guarantee of zero accidents in manufacturing. There are many variables. However, AI combined with timely human intervention can maximise the levels of safety across an operation. On the flip side, Gaurang Soni, Chief Technology Officer, Aavenir, believes that it is possible but it is quite far away; manufacturing has traditionally adapted to latest technologies to increase efficiency, production output & decrease overall cost. But it is yet to adopt latest technologies for workplace safety.
Deciding your AI budget
Plus, there’s one very important factor for companies to consider when turning to AI - the cost and budget factor. In spite of its benefits to workplace safety, AI is still a rising piece of high tech with a hefty price – an investment that needs careful consideration. “You can’t put a cost on human life,” avers Soni and continues, “Those who adopt AI early will benefit greatly as the tech industry is coming up with a lot of innovative solutions around it. This technology is here to stay, and we can quote numerous organisations that have already deployed it at their workplaces & have seen it make a difference immediately.”
Building AI systems is not only the cost of hardware and computers but very often, there is cost involved in ML and teaching these computers and so on. Therefore, it requires a lot of planning and investment. Kesavadas reveals that one way that manufacturing companies can decide on the investment is by really looking at the history of the productivity loss or the loss due to injury, etc and looking at the major hurdles or pain points in safety maintenance. “For example, let’s say there are 20 major accidents in a given year and if one can identify where the accidents were & how much it cost the company based on data, then one can calculate the ROI very easily if an AI-based CCTV or sensors on workers or any of those technologies are added,” he says.
AI is expensive – A myth
But is AI really that expensive? Ramachandran thinks not. He claims that it is a major myth that it is too expensive to apply AI for safety in manufacturing. In reality, organisations are increasingly using AI and are seeing ROI more quickly. “There is significant reduction in the costs of computing and storage while smart sensors and IoT act as enablers for gathering the necessary data. The beauty is that technologies today can run on a more simple device on the edge or on the cloud. Judicious application of AI and industrial analytics for identified use cases can quickly turn into a self-funding programme, allowing the organisation to scale to an enterprise level deployment over time,” he states. Likewise, Subramanian too, observes that AI being very expensive is a myth, which depends on use case to use case. Companies need to understand what the cost of safety is to them; it is not just something going wrong or some parts being damaged. It could be the cost of a human life.
AI skills in handy
One truly can’t put a cost on human life. And not only are the humans valuable themselves but the skills they impart in the industry. With AI growing so fast, there is going to be a demand for people from computer engineering, computer science and an understanding of how ML happens, etc. “The skills required to successfully deploy advanced AI technologies can be categorised into two major areas. Firstly, we clearly need a set of skills for solution providers in order to implement AI-based solutions. In this category, we need roles like data scientists, IT specialists, industry specific domain experts, business analysts and digital experts. Secondly, for deployment and adoption, we need customer expertise to define the use cases for business & safety processes,” Ramachandran divulges.
Adding to this, Soni coveyes, “One has to look forward and imagine that with the advancement of AI, we will have an abundance of data at our disposal. To make sense out of this huge amount of data, we will need data scientists to help give us a predictive output. Also, with predictive data at our disposal, there will be a huge opportunity to automate more processes and hence, automation skills.”
AI’s fight against COVID-19
So far, we know the role AI has been playing for the safety of the manufacturing industry and what needs to be done for the industry to benefit more deeply from it. We know what the manufacturing industry would need down the line in terms of skill-sets for AI. We know that through collaborative intelligence, humans and AI actively enhance each other’s complementary strengths. Nonetheless, what is the first thing that strikes us when we think of safety through AI in the times right now? Moreover, safety as a whole, in the times right now?
That’s right. We can’t look past the COVID-19 pandemic that has changed the complete functioning of industries and made safety a priority more than it perhaps will ever be. And AI hasn’t turned its back on the manufacturing industry whilst the COVID-19 outbreak has been giving everyone challenging times. In fact, India has the greatest advantage than any other country in the world right now, given the current scenario with COVID-19 that the world is facing. Lots of manufacturing units will move over to India with the amount of workforce we have, and this is the right time to take advantage of the same & embrace workplace safety with the help of AI.
“AI and ML are playing a key role in better understanding and addressing the COVID-19 crisis. ML technology enables computers to mimic human intelligence and ingest large volumes of data to quickly identify patterns & insights,” Parashivamurthy remarks and goes on, “Some of the main areas where AI is helping fight against COVID-19 are contactless screening of COVID-19 patients with symptoms by using AI chat bot, speeding up innovation in the field of vaccinations & treatments, helping to recognise patterns in X-ray images and enhancing the ability of radiologists to decide about COVID-19.”
Ramachandran goes on to further reveal, “Almost 50% of the businesses we speak to want to know more about how AI can be used to tackle the challenges posed by the pandemic. Increasingly, there is a need to run operations safely but with fewer human beings present at the site. Adoption of digital technologies & AI can greatly assist human beings to manage and operate many of the plant activities remotely.”
Putting AI to use
Through the advancement in technologies, AI has a huge potential to drastically reduce safety incidents and can provide greater safety as well as increase the visibility of problems before they turn into a crisis. The focus should be on embracing these technologies and putting them to use. Moreover, people must follow the warnings provided by the AI systems once they’re built, otherwise it won’t help. What industries have to understand is that AI requires a lot of coordination, planning and insight before they can actually find out what is an ideal solution that can reduce accidents inside the plant.