As debate rages over AI displacing jobs, history offers some insights

[By Peggy and Marco Lachmann-Anke under Creative Commons]

“It looks like WhatsApp is joining the WhatsApp moment in Indian banking!” – Nandan Nilekani. 

Nandan Nilekani was referring to WhatsApp’s reported plans to leverage the Indian government’s Unified Payments Interface (UPI) platform in its insanely popular chat platform in India to introduce peer-to-peer financial transactions. If true, this is expected to be a game changer in India’s accelerating journey to embrace digital transactions. The numbers say it all. WhatsApp had about 160 million active monthly users in India by the end of 2016. In contrast there were only about 28 million credit cards issued in India.

The second half of Nilekani’s tweet refers to an argument that he has been making for the past several months—that Indian banking is facing a WhatsApp moment, that the big banks today would lose their dominance and market share to new, smart and fast players, just like how big telecom players lost their dominance in messaging to WhatsApp, a company that employed about 50 people. The reason: mobile, cloud, data and a host of new digital technologies. The very same technologies are set to transform banking. Digital transactions have been picking up steam for some years now.

Demonetisation has given a fillip to digital transactions in India. The Indian government has set a target of 25 billion digital transactions in 2017-18. While this is an ambitious target, India seems to be on its way. According to the Reserve Bank of India (RBI), February 2017 saw about 763 million digital transactions. Urban Indians, and increasingly many in rural India as well, take these technologies in banking as a hygiene factor. The IT solutions that allow customers to transact with the bank directly are the front-end systems that hog the limelight. However, they are the proverbial tip of the iceberg. The heart and brain of retail banks is the core banking systems. These are IT solutions that help banks centrally automate all retail banking process like ATM transactions, debiting and crediting an account, changing a PIN, etc. Apart from the core banking systems, networks like NEFT, RTGS, IMPS, etc. have played an important role to facilitate self-serviced inter-bank transactions. 

While these exciting IT-led developments in digital transactions and digital banking promise to make our lives simpler and hassle free, other IT-led developments are expected to make some of our lives difficult. Artificial intelligence (AI), including robotic process automation (RPA) via chatbots, is likely to lead to job losses in millions. There are already murmurs that Indian jobs will be lost as well though Indian banks have been quick to deny this. So far there has not been any move to regulate the introduction of AI.  However, Bill Gates has suggested that governments should tax companies that use robots that take away human jobs. Predictably, this has triggered a debate. This may just be the beginning of more debates, and the government may be forced to take notice soon.    

The dawn of IT in Indian banking

History offers us insights into how to navigate these contexts. As Confucius famously said, “Study the past if you would define the future.” The dawn of IT in Indian banking from the 1950s to the 1980s teaches important lessons valid in today’s context as well.

First, the IT-driven business process outsourcing (BPO) model was popular even in the 1960s although it had a different name—service bureau. And IT-driven BPO, including in its latest avatar of chatbots, will always appeal to banks, especially for processes that are reasonably structured.

Second, the debate on computers replacing human workers is always a part of technology and IT evolution. The Indian banking sector has been debating this from the 1970s. There are no easy answers to questions on the socio-economic impact of IT replacing humans. A few parameters to consider while assessing the impact are national priorities—current and future, identifying alternative jobs for the displaced workforce, reskilling the displaced workforce, and whether IT adoption leads to a greater good for a majority, etc.

Third, security aspects associated with new IT will continue to be in the forefront. India will need the best-of-breed security-related processes and IT to confidently reassure all its citizens, from the savvy to the novice, that digital banking, digital transactions, and digital identities are safe.    

This sets the context for us to unravel answers to some questions: When was IT first used in Indian banks? Indian banks’ journey of IT adoption began in the 1950s. What milestones shaped the dawn of IT in Indian banking between the 1950s and 1980s? We dip into research done as part of the itihaasa history of Indian IT project to trace this extraordinary journey.

Our quest for the first computer used in an Indian bank takes us back in time to the late 1950s, a little after the Imperial Bank of India was nationalised and rechristened as State Bank of India (SBI) in 1955.  SBI was born with about 500 branches. An early directive to SBI (which incidentally, is as valid today as it was 60 years back) was to open more bank branches in every nook and corner of India.

As the number of branches increased post 1955, manual inter-bank reconciliation became more arduous. Inter-bank reconciliation is the matching of payments with receipts in banking transactions that involve more than one bank branch. In the late 1950s, it is believed that SBI obtained a unit record machine from ICT (International Computers and Tabulators) and installed it in Kolkata, then known as Calcutta. The ICT unit record machine is not an electronic computer but a mechanised computing machine. The unit record machine was hardwired for a particular function. A set of metal rods in a connection box had to be soldered to achieve the functionality. This technology was originally from Powers-Samas Accounting Machine Ltd, which merged with British Tabulating Machine Company to form ICT. Interestingly, the first electronic computer in India was the HEC 2M installed at the Indian Statistical Institute in Kolkata in 1955. It was also a product of the British Tabulating Machine Company.

Soon, like any IT system today, SBI’s demand for computing power outgrew the ICT unit record machine. Sometime in the early 1960s, SBI migrated to an IBM 1401 electronic computer. The IBM 1401 was the workhorse of computing in India in the 1960s and early 1970s. Close to a hundred IBM 1401s were in use in India by 1970. The story of the rise of IBM 1401s in India and its subsequent connection with Indian banking is interesting.

In 1959, IBM started a factory in Mumbai (then Bombay) to manufacture key punch machines. IBM was exporting many of these key punch machines manufactured in Mumbai and earning forex. According to the licence IBM had from the Indian government, it could import used computers valued up to 80% forex that it earned. It is said that IBM used to import discarded IBM 1401s at a very low inter-company transfer price. This was because the IBM 1401s in markets like the US were increasingly getting replaced by the IBM 360 and other newer computer systems by the mid-60s.

These IBM 1401s were refurbished in the Mumbai factory, and leased to Indian organisations at what was considered to be a steep fee. For an annual fee of about Rs 10 lakh, IBM would lease the IBM 1401, provide maintenance, and the services of an IBM systems engineer who could design the system as per client needs. In many cases, the dominant business model of IBM in India at that time was that of a service bureau. In today’s context, it may as well be thought of as the IT-leveraged BPO model. Even its detractors admit that IBM provided excellent service to its Indian customers. The IBM 1401 went on to play an important role in the computerisation of Indian banks.

Apart from SBI and IBM, there was a new Indian business entity that used IBM 1401s to service banks.

The Tata group had set up an internal Tata Computer Centre group in the mid-60s to facilitate the use of computers in different Tata group companies. This group became a division of Tata Sons and was renamed as Tata Consultancy Services (TCS) in 1968. By this time, TCS had two IBM 1401s on lease and installed in its swanky office in Nirmal building, Nariman Point (then a developing area reclaimed from the sea in south Mumbai).

According to Dr. Lalit Kanodia (a senior member of the Tata Computer Centre and TCS in the mid-60s and founder of Datamatics), one of the early large non-Tata group clients of TCS was the Central Bank of India. TCS with its computer expertise worked on inter-branch reconciliation for the Central Bank of India using a service bureau model. Unlike today, programming was a two-step process on the IBM 1401. The programs were written by a few programmers. TCS employees of this era, like S Mahalingam (who spent 42 years in TCS and went on to become its CFO), mention there were less than 25 programmers who were recruited from the best universities in India and abroad. These programmers were fondly referred to as the “head” of TCS. They used an assembly language called Autocoder to program the IBM 1401. Data punch operators did the work of punching the cards that contained client data, and controllers operated the computer—i.e. they fed the input and obtained the output. There were about 250 data punch operators and controllers in TCS, and were fondly referred to as the “tail”. To place it in perspective, by 1969-70 TCS had less than 300 employees. By December 2016, TCS had about 378,000 employees and is among the largest employers in the private sector.

While Indian banking was taking baby steps in adopting IT, the late ’60s and early ’70s was a tumultuous time for India. We had fought wars in 1967 and 1971. Socialistic policies were dominant. All private sector banks were nationalised in 1969. The public sector was seen as the torch bearer of industrialisation, including computerisation. Forex was always at a premium and the rupee was not convertible. Against this backdrop, in 1969 the government set up a committee chaired by R Venkataraman (who went on to become the President of India) to study the role of automation on employment in India. VM Dandekar took over from R Venkataraman as the committee’s chair and released a report in 1972. The Dandekar Committee Report prescribed controls for using computers in government and industry. The committee’s recommendations made it mandatory to obtain the formal agreement of labour before introducing computers. This ensured that rapid bank computerisation took a back seat.

Indian IT doyens like Prof. V Rajaraman (former faculty at IIT Kanpur and Indian Institute of Science and one of the founding fathers of computer science education in India) point out that the IBM 1401 was partially to blame for this position of the Dandekar Committee Report. While this seems counter-intuitive at first glance, the logic is compelling. The Dandekar Committee got a sense from the white collar labour familiar with computerisation that the IBM 1401 was perceived more as a labour saving device rather than a productivity enhancing device. The Parliament’s Public Accounts Committee in the mid-70s also echoed these sentiments. While they agreed that in the long run computers are useful to increase productivity, they felt that the use of computers in India was detrimental given India’s large-scale unemployment. This was the proverbial last nail in the coffin of bank computerisation in the ’70s. In the present context, this has a parallel with current emerging technologies like RPA and AI threatening to replace humans in many business processes.

In 1978, licences were given to Indian business entities to manufacture and sell small minicomputers, breaking the monopoly of the public sector Electronics Corporation of India Ltd (ECIL) that manufactured larger computers. UNIX (especially the UC Berkeley version of UNIX) made its debut in India as a preferred operating system in these minicomputers since it was a multiuser operating system. Since minicomputers were a new technology, the utility was not immediately evident to many Indian businesses. It was perceived as an expensive asset, and UNIX provided the minicomputers the flexibility to be shared among different functional organisations in a company. As we will discover shortly, minicomputers and UNIX played an important role in Indian banking.

The early ’80s gave rise to a new hope. India had a young Prime Minister Rajiv Gandhi and his advisors like Dr. N Seshagiri who believed in the power of technology, and computers in particular. The topic of bank computerisation came to the limelight again. In 1984, a committee set up by the RBI and chaired by Dr. C Rangarajan (who was then deputy governor of RBI and later became governor of RBI) was set up to present a roadmap for computerisation of the banking industry for the five year period 1984 to ’89. The most important recommendation was that banks should set up electronic data processing (EDP) cells a.k.a. IT organisation of today, and use computers for back-office work. There was an interesting observation on the use of technology partly motivated by the developments in the late ’70s. The Rangarajan Committee recommended the use of UNIX as the operating system in banks. This technical recommendation was futuristic and helped Indian banking to leapfrog mainframe technology in favour of the newer minicomputer technology. The rationale was that mainframe technology was not required for the volume of work at an Indian bank branch level, and the emerging modern computing technology in the form of minicomputers was adequate.

At the same time, the mandarins in RBI were pragmatic. They suggested the use of IBM mainframes for computerising cheque clearing operations—each of the four original metros of India got a mainframe.  This was required to improve the efficiency of cheque clearing that often took a few weeks in that era, and a must have infrastructure to motivate Indians to transition from cash to cheques. This is similar to the current context where infrastructure like UPI is helping the transition from cash and cheques to mobile. IT adoption in the mid-80s was neither instantaneous nor smooth. The trade unions were quick to respond, and they declared 1984 as “anti-computerisation” year. As the story goes, the Rangarajan Committee was aware of the trade unions’ stand. The word computer was rarely mentioned. The committee was officially called the Committee on Mechanisation in the Banking Industry (1984). And computers were called “Advanced Ledger Posting Machines (ALPMs)!

While bank branch automation was one aspect of computerisation being conceived, the RBI also realised the importance of networking banks within the country and joining international banking networks. In 1987, the RBI Committee on Communication Network for Banks and SWIFT Implementation (SWIFT is short for the Society for Worldwide Interbank Financial Telecommunication) recommended setting up a X.25 packet switching network—BANKNET—to facilitate inter-bank fund transfers within India. The ingenuity of the committee was evident when they suggested that the four IBM mainframes in the four metros used for cheque clearing operations could be used in the banking hours as hubs for BANKNET in Phase 1 of its implementation. And the IBM mainframes will be used for cheque clearing operations outside of banking hours. The committee also recommended that India should join the international SWIFT network, and BANKNET should use the same security and messaging protocols like in SWIFT. In the current context, we are increasingly debating about the appropriate security and privacy standards India should adopt for providing the best and safe experience in digital banking and digital transactions.

While the 1984 RBI Committee on Mechanisation in the Banking Industry hogs the limelight, the RBI Committee on Computerisation in Banks of 1988 (again chaired by Dr. C Rangarajan) is often missed out. This committee was to recommend the roadmap of bank computerisation for the five years 1989 to ’94. Interestingly this probably is the first time the phrase “computerisation in banks” was used without an apology. The reason for the positive shift in perception of computers from 1984 to 1988 is often attributed to the fact that “mechanisation” had led to productivity benefits without job losses. While this is true, there was another interesting and unexpected related narrative as well.   

By 1988, the general Indian public perception of computerisation was extremely positive. This had little to do with the use of computers in banking. The year 1986 witnessed the inauguration of computerised railway passenger ticketing in India. The benefit to an average Indian citizen was huge. For starters, one did not have to apply for a day’s leave or lose a day’s pay for making a railway reservation. This was an eye opener for the government and policy makers on the potential of computerisation. They realised that the pull for computerisation from the citizens could compensate the resistance from trade unions.

The RBI Committee on Computerisation in Banks had a recommendation that paved the way for digital banking. It recommended the introduction of credit cards with widespread acceptance across merchant establishments in India to reduce the load on cash and cheque transactions. This committee also recommended setting up of a network of ATMs in Mumbai as a pilot. While the banks had an arduous journey in the late ’80s to adopt IT, these initiatives provided the base for core banking and electronic fund transfer that came subsequently. And in turn, these were the basis for the evolution of IT in Indian banking we have witnessed since the Internet era.   

As Indian banking enters a new phase—driven by disruptive technologies, different customer expectations, and a new kind of competitive landscape—it would help to remember the lessons from history. Business logic always wins, and dealing with the impact of technology is not an easy business. 

(The views are personal.)