Essay on Rise of Artificial Intelligence: Jobless Future or Reskilling Opportunity

“Rise of Artificial Intelligence: the threat of jobless future or better job opportunities through reskilling and upskilling” appeared as Topic 8 of Section B in the UPSC Civil Services Mains Essay paper held on 20 September 2019. One long, deliberately loaded sentence on the question paper. One hundred and twenty-five marks if you handle it well; a wasted hour if you do not. This guide breaks the topic down the way it should be tackled in the hall — first the framing, then the angles, then the time map, then the paragraph-by-paragraph plan, and finally a complete 1,200-word model essay you can study, dissect and adapt.

When this was asked and what UPSC is really testing

The topic was set in UPSC CSE Mains 2019, Essay paper, Section B, Topic 8. Three hours, two essays, one from each section, roughly 1,000 to 1,200 words each, 125 marks per essay. Unlike a pure philosophical prompt, this one is policy-loaded: it forces you to take a position on technology, labour and the State at the same time. The phrasing itself is a trap — it offers two alternatives and dares you to pick one. The mark-pulling answer refuses the false binary and constructs a third path.

UPSC is checking four things at once. One, can you read a technology question without writing a Wikipedia entry on neural networks. Two, can you bring economic theory, historical parallel and Indian policy together without sounding like three different students. Three, can you handle a counter-argument honestly — admit the transition pain even while you argue for the upside. Four, can you write 1,100 disciplined words that examine a public policy choice rather than 1,500 loose words that celebrate or denounce a technology. The aspirant who handles all four scores in the 130s. The one who only describes “what AI is” scores in the 80s.

Five angles that unlock the artificial intelligence essay

The trap with a binary prompt is to pick one side and ride it for 1,100 words. The mark-pulling answer instead identifies several lenses, names them clearly, and weaves them. Here are the five most defensible readings of the topic. You do not need all five — three, treated well, is the sweet spot. But you should know all five so your choice is conscious.

1. The historical-parallel angle — the steam engine displaced handloom weavers but created factory workers; electricity wiped out lamplighters but built the appliance industry; the personal computer hollowed out typing pools but birthed the IT services sector. Each wave destroyed jobs and created different jobs, on different time scales. The question is whether AI is the next wave, or whether cognitive automation is qualitatively different from muscle automation.
2. The two-camps angle — the pessimist school, associated with Daron Acemoglu and Pascual Restrepo, argues that AI is a “so-so” technology that displaces workers faster than it creates new productive tasks. Klaus Schwab’s “Fourth Industrial Revolution” warns of unprecedented churn. The optimist school, associated with Erik Brynjolfsson and Andrew McAfee in The Second Machine Age, argues that entirely new categories of work will emerge — as they always have.
3. The India-specific angle — twelve million Indians enter the workforce every year; the demographic dividend window closes around 2055; a large share of formal employment sits in IT services, BPO and call-centre work — precisely the layers most exposed to generative AI. India cannot afford a jobless future, and cannot wait a generation for “new jobs” to emerge.
4. The sectoral-shift angle — radiology, legal due diligence, customer support, basic coding, precision agriculture, logistics routing — each sector is being reshaped on its own timeline. Some white-collar tasks are now more exposed than blue-collar ones, reversing the usual story of automation. Naming three sectors precisely beats waving at “AI in every industry”.
5. The institutional-response angle — IndiaAI Mission (2024), NITI Aayog’s National Strategy for AI (2018), Skill India, PMKVY, NSDC, NEP 2020’s multidisciplinary push, Singapore’s SkillsFuture credits, the European AI Act. The question is no longer whether AI will disrupt labour, but what public infrastructure cushions the transition.

A strong essay picks angles 1, 3 and 5 as its backbone (history, India, policy), uses 2 to stage an honest debate, and 4 to ground the abstraction in concrete sectors. That gives the essay rhythm — context, complication, resolution — instead of a single straight line.

A time map for the 1,500-second window

One essay deserves about 75 to 90 minutes inside the three-hour paper. Spending more on Essay 1 starves Essay 2. The single biggest source of below-100 essay scores is poor time discipline — beautiful introductions, panicked conclusions. Use a rough breakdown like this:

Notice what is not on the list: rewriting the introduction halfway, hunting for the perfect quote, fancy diagrams, decorative underlining. None of that earns marks. Discipline does.

What to add — and what to avoid at all costs

The essay paper rewards what most aspirants under-do and punishes what most aspirants over-do. Memorise this list before you walk into the hall.

Add liberally

• One precise thesis, stated by the end of paragraph two and not abandoned. “AI will not deliver a jobless future, but it will deliver an unevenly distributed one; the role of the State is to convert that unevenness into mobility through public reskilling infrastructure.”
• Concrete examples across three or four domains — one historical (steam engine, electricity, the 1990s PC wave), one Indian (IT services, BPO, IndiaAI Mission), one scientific or sectoral (radiology, precision farming, GitHub Copilot) and one philosophical (the Gita on action without attachment, Gandhi on machinery serving humans).
• Two or three short, named references — Brynjolfsson and McAfee on the Second Machine Age, Acemoglu on so-so technologies, Klaus Schwab on the Fourth Industrial Revolution. One per major section is enough.
• One Indian philosophical anchor. Gandhi’s distinction between machinery that serves man and machinery that enslaves him, or the Upanishadic principle that knowledge which liberates is vidya and knowledge which binds is avidya. An Indian anchor in a tech essay full of Western names stands out.
• Counter-arguments handled honestly. “Reskilling can become a polite euphemism for fend-for-yourself…” — then resolved in two lines. Acknowledging the human cost earns more marks than ignoring it.
• A clean conclusion that returns to the opening image — not a fresh argument, not a new statistic.

Avoid — even when tempted

• Defining what AI is in the first paragraph. “Artificial intelligence is a branch of computer science that…” — this signals that you have nothing to add. Open with a scene from a factory floor or a hospital reading room.
• Drifting into a pure tech-bro celebration. An essay that only lists what AI can now do reads like a brochure, not an essay. The paper rewards judgement, not enthusiasm.
• Unsourced statistics. “85% of Indian jobs will be automated by 2030” — if you cannot name the source, do not write the number. Examiners cross-mark this.
• Hot-button political examples. The Essay paper is graded by humans across the ideological spectrum. Naming current parties, ministers or sitting judges introduces avoidable risk. Use the institution, not the personality.
• Decorative scholar-dropping. Citing Yuval Harari, Nick Bostrom and Ray Kurzweil in one paragraph to look read. Examiners spot performance citations instantly. One scholar, used well, beats four named in passing.
• Moralising about the future. “We must embrace technology with caution” reads like a school speech. The essay paper rewards examination, not exhortation.

A paragraph-by-paragraph blueprint

Twelve paragraphs at roughly 90 words each gets you to 1,080. Fifteen at 75 each gets you to 1,125. Either works. What follows is a 13-paragraph plan that maps cleanly onto the model essay below. Each line is what that paragraph is doing, not what it is saying.

1. Opening scene — a radiology reading room or a Bengaluru call-centre floor at midnight. Physical, specific, no jargon. No mention of the topic yet.
2. The pivot to thesis — name the binary in the topic, refuse it, state the third position in one sentence.
3. Define the terms carefully — what “AI” means in 2019 versus today, what “jobless future” actually claims, what “reskilling” promises and obscures.
4. Historical parallel — steam engine, electricity, the personal computer. Each wave destroyed and created. The pattern, and the caveat that cognitive automation may differ.
5. The two camps — Acemoglu and Schwab on one side, Brynjolfsson and McAfee on the other. Stage the debate, do not pick yet.
6. India’s specific exposure — demographic dividend, twelve million entrants annually, IT-services concentration, the BPO question.
7. Sectoral picture — radiology, legal due diligence, agriculture, coding. White-collar exposure rising; new sectoral lines being drawn.
8. New categories of work — AI ethicists, prompt engineers, data labellers, the care economy. Honest about the scale gap with what is being displaced.
9. The counter-argument — reskilling as euphemism; older workers stranded; the gap between policy slogans and lived transition.
10. Indian policy response — IndiaAI Mission, NITI Aayog’s National AI Strategy, NEP 2020’s multidisciplinary education, Skill India and PMKVY’s record.
11. Comparative angle — Singapore’s SkillsFuture credits, Germany’s dual-track apprenticeship, what India can learn without copying.
12. The Indian philosophical anchor — Gandhi on machinery in service of man, the Gita on right action without attachment to fruit. The ethical frame for the policy choice.
13. Closing image — return to the radiologist or the call-centre worker, this time with the third path drawn around them.

How to make an examiner stop and read

An essay examiner reads several hundred scripts in a sitting. The first paragraph decides whether they read the second with attention or with autopilot. Four small habits separate the essays that get attention from the ones that get skimmed.

• Open with a scene, not a definition. A radiologist watching a model flag a tumour, a weaver in 1820s Lancashire, a Bengaluru data labeller on a night shift. Concrete images cost no marks and earn attention.
• Use the topic phrase itself two or three times across the essay. Once in the thesis, once at the turn, once at the close. It signals discipline and prevents drift.
• Vary sentence length. A short sentence after three long ones lands. Examiners feel rhythm before they parse meaning.
• End paragraphs with the takeaway, not with examples. Place the example mid-paragraph. Let the last sentence be the thought, so the reader carries it into the next paragraph.

The complete essay — “Rise of Artificial Intelligence: the threat of jobless future or better job opportunities through reskilling and upskilling”

What follows is a 1,200-word model essay built on the blueprint above. Read it twice — once for the argument, once for the moves. The moves are what you can carry into your own essay; the argument is one of many you could make.

In a darkened reading room at a Bengaluru hospital, a radiologist watches two screens. The left shows a chest scan. The right shows the same scan with three regions outlined in red by a software model trained on a million images. She accepts two flags and overrules the third. The report goes out with her signature. A decade ago she would have done the work alone. A decade from now, the question is whether she will be doing it at all — and if she is, what she will have learned to do that the model cannot.

The topic offered to UPSC candidates in 2019 forces a choice between two futures: a jobless one delivered by artificial intelligence, or a better one delivered by reskilling and upskilling. The honest answer is that neither is right. The rise of artificial intelligence will not produce a single future at all. It will produce a steeply uneven one, in which the upside and the downside fall on different people in different places at different speeds. The work of public policy — and the test of this essay — is to refuse the binary and ask instead what institutions a republic must build so that the gains spread and the losses do not concentrate.

It helps to be precise. “Artificial intelligence” in 2019 mostly meant narrow systems doing one task well; today it means generative models writing prose, code and images at near-human plausibility. “Jobless future” is shorthand for the claim that this round of automation displaces work faster than it creates it. “Reskilling and upskilling” is the policy reply that workers can be retrained into new categories. Each term carries an argument inside it; the strongest essay handles all three with care.

History does not settle the matter, but it sets the terms. The steam engine destroyed the handloom weaver and built the factory worker. Electricity ended the lamplighter’s trade and created the appliance industry. The personal computer hollowed out typing pools and travel agents through the 1990s — and produced an IT services industry on whose shoulders India built a middle class. Each wave displaced labour and, on a longer timescale, created different and often more productive labour. The caveat is real: every previous wave automated muscle and routine. Artificial intelligence is the first that automates cognitive judgement. Whether the pattern will hold is the question on which honest scholars disagree.

Two camps dominate that disagreement. Daron Acemoglu and Pascual Restrepo argue that AI today is largely a “so-so” technology that displaces workers without raising productivity enough to create offsetting tasks; Klaus Schwab’s account of a Fourth Industrial Revolution warns of churn at unprecedented scale. Against them, Erik Brynjolfsson and Andrew McAfee in The Second Machine Age argue that wholly new categories of work will emerge as they did in every previous wave, and that the constraint is institutional rather than technological. The disagreement is not academic; it maps onto two very different policy bets.

India faces that choice with a clock running. Roughly twelve million Indians enter the workforce each year; the demographic dividend window closes around 2055. A disproportionate share of formal employment sits in IT services, business process outsourcing and call-centre work — precisely the layers most exposed to generative AI. A jobless future is not a possibility India can afford; nor can the republic wait a generation for new categories of work to settle. The pace of policy must match the pace of the technology, or the dividend becomes a burden.

The sectoral picture is already visible. Radiology and pathology now use models that match human accuracy on narrow diagnostic tasks. Legal due diligence, once a junior associate’s monopoly, is being compressed by document-review systems. Call centres are being thinned by voice agents. In agriculture, precision-farming sensors and satellite-based yield estimates are quietly altering extension services. In software itself, tools such as GitHub Copilot have reordered what a first-year coder does in a day. The familiar reassurance — that automation hurts the blue-collar worker and rewards the white-collar — has reversed.

New categories of work are emerging in the same breath. The AI ethicist, the prompt engineer, the data labeller, the AI auditor, the human-in-the-loop reviewer — none of these titles existed in meaningful numbers a decade ago. Beyond them lies the wider care economy: nursing, early-childhood education, eldercare, skilled trades, hospitality. None of these can be replaced by a model that has no body and no felt life. The honest difficulty is that the new categories are still smaller than the displaced ones, and that the people losing the old jobs are rarely the same people gaining the new.

This is where the counter-argument must be heard. “Reskilling and upskilling” can quietly become a euphemism for fend-for-yourself, moving the cost of structural change from the firm and the State onto the worker. A forty-five-year-old call-centre supervisor in a Tier-2 city cannot retrain into prompt engineering by attending a weekend workshop. Transition pain is real, falls hardest on older workers and on women re-entering the workforce, and pretending otherwise hollows the policy out before it begins.

The Indian state has begun to assemble its response. NITI Aayog’s National Strategy for Artificial Intelligence (2018), the IndiaAI Mission (2024), the Skill India ecosystem, the Pradhan Mantri Kaushal Vikas Yojana, the National Skill Development Corporation and the National Education Policy (2020) form the bones of a policy architecture. What is missing is flesh: a portable lifelong-learning credit that follows the worker across employers, social security that travels with gig and transitioning workers, a serious conversation on a minimum guaranteed income, and an AI governance framework that protects against the worst harms without freezing the upside.

India does not need to invent every piece. Singapore’s SkillsFuture gives every adult citizen a credit spendable at accredited providers across a lifetime; Germany’s dual-track apprenticeship keeps employers in the room with educators; the European Union’s AI Act tries to bind risk to oversight without choking innovation. Importing models is easy and usually a mistake; adapting them to Indian scale and federal complexity is harder and usually correct. STEAM education in schools, multidisciplinary universities under NEP 2020, and serious investment in vocational dignity must arrive together.

Beneath the policy lies an older question. Gandhi distinguished between machinery that serves human ends and machinery that enslaves humans to its rhythms; he opposed not the loom but the mill that turned the weaver into its servant. The Upanishads named two kinds of knowledge — vidya that liberates and avidya that binds. Artificial intelligence is neither in itself; it becomes one or the other depending on the institutions a society places around it. A republic that treats workers as costs to be optimised will produce one outcome; a republic that treats them as citizens whose mobility is the State’s responsibility will produce another.

Return to the reading room with which we began. The radiologist’s signature still goes on the report. What has changed is what she is doing — she is no longer the only pair of eyes; she is the judgement that holds the model accountable. Multiply that scene across a hundred sectors and the shape of the choice is clear. The rise of artificial intelligence will neither produce a jobless future nor automatically deliver better jobs. It will deliver the kind of country we build around it.

Word count: approximately 1,200.

How to use this model essay

Do not memorise it. Memorised essays read like memorised essays, and examiners spot them in two paragraphs. Use the model the way a chess student uses a master game: study the opening scene, the refusal of the binary in paragraph two, the way historical parallel hands off to the two-camps debate, the placement of the Indian anchor, the way the counter-argument is brought up and then closed. Then take a different technology essay topic — “Technology cannot replace manpower” or “Science is a boon or bane” — and write your own essay using the same blueprint. Repeat that exercise eight to ten times and the structure becomes muscle memory. On the day of the exam, the only thing you should have to think about is the topic itself.