It is time to ask all over again if AI is killing software engineering. Not because things are not clear but because some statements from industry leaders offer more confusion than clarity. It is wise to keep in mind futurist Roy Amara’s Law that states that we tend to overestimate the effect of a technology in the short run and underestimate its effect in the long run.

Now, what does this mean truly in the here and now? That question popped up in my mind after Tata Group Chairman N. Chandrasekaran, who also heads Tata Consultancy Services (TCS), India’s largest software services exporter, told TCS shareholders that hiring would slow as the company moves towards having an equal number of employees and AI agents in its workforce. AI agents, he said, may number about half a million in three years, roughly equal to the company’s current workforce.

That statement is a tad disturbing because it comes a year after the company cut its workforce by about 8,000 over the past year, slightly less than the planned 12,000. Clearly, TCS is signalling that job growth may not match industry growth. That is not so good news for fresh software graduates. There is more to be concerned about because a country like India is in the throes of a demographic transition. More than 65% of India’s population, or in excess of 80 crore people, are below the age of 35.  

Needed: A Roadmap

Information technology has been the passport to success for millions of youths over the past three decades. What we therefore need is not a broad-brush statement but a roadmap, or at least a working model of one, from both companies like TCS and the government. India’s IT industry is worth US$ 315 billion in revenues and has been a key source of exports and foreign exchange reserves. With the recent weakening of the Indian rupee against the US dollar, not to speak of President Donald Trump’s whimsical crackdown on work permits like H1B visas for Indians, things are getting more complicated.

Let us talk a bit about rickshaws to get a hang of what the future might look like. As late as 1991, I saw my first hand-rickshaw in adult memory in Kolkata (then Calcutta), even as I had my first Indian metro ride in the same city. The visit brought back memories of actor Balraj Sahni as a suffering farmer turning a Calcutta hand-rickshaw puller in ‘Do Bigha Zameen’. That is the stuff of social and cinematic history. But I also recall that, circa 1973, DMK leader and Tamil Nadu Chief Minister Muthuvel Karunanidhi presented cycle rickshaws in Chennai (then Madras) to liberate hand-rickshaw pullers from hard labour.  Unemployment and convenience are two sides of the same tech coin. It is time to consider AI with a similar outlook.

In fact, the TCS chairman spelt out AI-linked opportunities as well at the company’s annual meeting.  He mentioned potential roles for TCS in clients upgrading their legacy technology infrastructure, data architectures, and outdated software to make them all AI-ready -- and then redesigning business operations through AI, re-imagining everything from supply chains to customer engagement, besides sharpening AI governance and management through training, monitoring, and securing AI-linked systems while keeping costs efficient.

Technological Climate Change

You can call it a new phase in technological climate change.

Decades ago, when Steve Jobs ushered in the desktop computer revolution, he described the computer as a ‘bicycle for our minds’. AI makes that mind machine fly, like a supersonic jet.  In a way, I liken the journey from printed paper to computing and then AI to the shift from hand-rickshaws to cycle rickshaws and then their auto version.

What we need to understand therefore is that just as airports that did not exist when the bicycle was invented emerged later, there will be an AI equivalent of airports and an entire ecosystem supporting it: Think of AI versions of airport security, stewards and stewardesses, maintenance engineers, taxi cabs, runways… the works. Those are the kind of things the TCS chief spoke of.

However, what concerns me is that AI is being described in linear terms as something that is happening smoothly within a country or an industry – while what in reality will happen is a rickety across-the-board revolution with ups and downs.

Companies will no doubt retrain engineers, but in a code-on-tap world. True, software apps can now be developed fast and cheap by those who know little or no code. But we will see a new set of intermediaries or old ones reinventing the way TCS has outlined.

Think of it like income tax returns. Theoretically, we can file our own income tax returns online, but many - if not most - of us need chartered accountants or tax-return preparers (TRPs) to make things easier because there is a world of difference between theory and practice.

Here’s where we need to understand that the journey from engineers to engineers using AI agents like “mind rickshaws” is not going to be quick, smooth, or easy. At each step, it would require a new set of professionals or old ones reskilling to run new systems, often involving inter-disciplinary thinking. This zigzag path is often underestimated by casual observers and stock market analysts. For revenues to match technological leaps, the zig-zag path has to be reasonably mapped or visualised.

Weather-Forecast Equivalent

That is precisely why we will need the weather forecast equivalent of AI adoption. India’s AI-aware ecosystem now includes universities and schools. Industry bodies like the National Association of Software and Service Companies (Nasscom) and government agencies need to sit together so that the awareness fructifies as a well-choreographed shift to new technologies. Forecasts are one of the biggest enablers of roadmaps so that costs, efforts, and outcomes are reasonably well synchronised.

Just as we decide to carry umbrellas or overcoats depending on what weather forecasts say, AI needs a forecasting engine to steer investments in education, adoption and security. It is not an easy task. Nor are weather forecasts. As we know well, monsoon predictions made months in advance guide us, even though they are not guaranteed to succeed. But history tells us they can help us manage risks in farming.

The government and industry need to work together so that irrational fears are replaced by workable concerns in a post-AI universe. AI-centric forecasts can be a great help in the process.

(The writer is a senior journalist covering a diverse range of subjects, including economy, technology, and politics. Views expressed are personal.)