Why the Indian Railways will face an AI conundrum

A majority of the jobs offered by Indian Railways in its recent recruitment drive, for which it received 25 million applications, can be automated with the help of AI and robotics

HYDERABAD: If you are among the tribe that believes that artificial intelligence (AI), robotics and automation will result in replacing jobs done by humans in India, you may wonder why the Indian Railways is preparing to recruit 90,000 people in what is touted as the world’s largest recruitment drive.

A majority of these jobs are for posts such as track maintainer, gangman, pointsman, switchman, helper and porter. The Indian Railways has received 18.9 million applications for 62,907 positions and 4.755 million applications for 26,500 positions of Automotive Loco Pilot. In fact, the Railways received a total of 25 million applications for these 90,000 jobs, most of which can, and will eventually, be automated with the help of AI and robotics.

The Indian Railways, for instance, can use automatic railway track cleaning systems, the Internet of Things (IoT) with multiple sensors to monitor employees and tracks, drones to monitor and scan tracks for faults and pilferage, and prevent potential derailments. The UK is already leading the way for robot train maintenance with university-led initiatives.

Train maintenance is typically performed in difficult working conditions and often under time pressure. It helps reduce human error in such high pressure and complex tasks.

Further, an automated train dispatching system can observe past scheduling decisions and outcomes to make better decisions with the help of AI (machine learning and deep learning). Besides, robotic welding increases the life of the tracks and crossing, making the process not only time effective, but cost effective as well. The repair process takes less than a day and there is no need to block traffic. Best of all, robots can work day and night without needing unions or rest.

The millions of job applicants must be mindful of the fact that technology has been pervading the Indian Railways for quite some time. On 8 July 2014, for instance, the railway budget had proposed bullet trains, bio-toilets, ultrasonic fixes for railway tracks, GIS (geographic information system) mapping, digitization of railway land, Wi-Fi connectivity at select stations and in trains, logistics support for e-commerce companies and going paperless in five years.

In his 2014-15 railway budget speech, the then railway minister Suresh Prabhu had unveiled an information technology (IT) vision for the railways, which included online information on the latest berth availability on running trains and an integrated mobile application including a station navigation system, besides customer-friendly freight movement initiatives such as introduction of barcoded/RFID (radio frequency identification) tracking of parcels and freight wagons, automated parcel warehouses and a customer relationship management system.

We do not have a separate Railway Budget any longer, but the government announced in the Union Budget 2018 that major steps include increasing use of technology like ‘‘Fog Safe’’ and a ‘‘Train Protection and Warning System’’. Moreover, a decision has been taken to eliminate 4267 unmanned level crossings in the broad gauge network in the next two years. Prime Minister Narendra Modi also doubled the allocation to Digital India programme to Rs3,073 crore in 2018-19.

Not to forget that the foundation for the Mumbai-Ahmedabad bullet train project, India’s first high speed rail project was laid on 14 September, 2017. There is a lot of automation in our metro and bullet trains. Since human drivers are unable to read signals at very high speeds, bullet trains have a different kind of speed-control system, known as Automatic Train Control (ATC), which allows for speed information to be transmitted along the track and received by a signal attached to the driver’s seat.

Further, magnetic levitation (which explains why we call it “Maglev”) for bullet trains is achieved through the use of an electrodynamic suspension system, or EDS. And you surely can’t have traditional track maintenance for maglevs—these are clearly software-driven tasks with minimal human intervention.

The millions who have applied for these thousands of routine jobs have also done so at a time when Piyush Goyal, Union Minister for Railways and Coal, said at a recent AI event in New Delhi that “AI is about creating trains with brains…(that) AI can transform Indian Railways in terms of safety, passenger amenities, better revenues, growth and efficiency”, and that it can “…be harnessed to find digital innovations for better customer interface and better service delivery”. He cited the example of how Google Maps’ data, for instance, could be intelligently used to estimate the time taken to travel from one point to the other.

To be sure, he did comfort the audience saying that AI should be used just as a tool and not to replace human jobs, an extremely sensitive subject in a developing country like India.

A 2017 December McKinsey Global Institute report notes that by 2030, 9% of today’s jobs in India will be done by machines, 16% in China, 23% in the US and 24% in Germany. The global average is 15%.

Martin Ford contends in his book ‘The Rise of the Robots: Technology and the Threat of Mass Employment’, that while all jobs are at risk of automation, it is the “routine” and “predictable” jobs that will be impacted most. For instance, it’s not hard to imagine that many drivers would be rendered redundant if driverless cars and trucks eventually go mainstream. Similarly, AI- and drone-driven surveillance systems can drastically reduce the number of security guards. Of course, they will raise questions of privacy, which policy makers will have to take cognizance of .

Ford cites a 2013 study by the University of Oxford’s Martin School of over 700 job types and concluded that nearly 50% of US jobs will ultimately be susceptible to full machine automation. Ford contends that robotics and advanced self-service technologies will primarily threaten lower-wage jobs that require modest levels of education and training. Automated vehicles or construction-scale 3D printers may eventually destroy millions of jobs.

On the flip side, Guido Jouret, chief digital officer (CDO) of ABB Ltd, pointed out in an interview to Mint that more automation also “creates better jobs, and ultimately even more jobs”. “Look at the three countries with the highest level of the adoption of robotics—Germany, Japan and South Korea. They also have some of the lowest rates of unemployment,” he pointed out.

The McKinsey report, too, points out that while automation will displace some workers and transform occupations, “we also know that new and additional work will be created in the next decade and beyond”. It does caution, though, that “what is less clear is how job growth net of automation will vary by occupation, and under what conditions there will be enough new jobs to offset the work that is lost as robotics, artificial intelligence, and other technologies assume a greater role in the workplace”.

AI will undoutedly introduce many benefits and help the Indian Railways become more efficient and productive since the latter has humoungous amounts of data from which meaningful business intelligence can be derived. Simultaneously, though, it will have to give serious thought to retraining its already huge 1.3 million workforce for jobs of the future.