Long-Term Effects of Artificial Intelligence on Employment and Wages in India
Using a combination of simulated microdata analysis and official aggregate statistics, this paper concludes that while AI has not caused aggregate job destruction in India, it is driving a significant skill and formality divide by increasing wages for graduates while penalizing non-graduates with lower pay and higher informal employment.
Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer
The Great Digital Shift: Why Your Job Might Change Before You Even Start
Imagine the economy as a giant, bustling kitchen. For decades, the recipe for success was simple: learn a skill, show up, and get paid. But recently, a new kind of sous-chef has walked into the kitchen: Artificial Intelligence (AI). This isn't a robot that just chops vegetables; it's a smart assistant that can write recipes, manage the inventory, and even talk to customers. The big question economists are asking is: Is this new assistant going to help everyone cook better and earn more, or is it going to kick some workers out of the kitchen entirely?
To understand this, we need to know two things. First, "skill polarization" is like a seesaw. When a new tool arrives, it often helps the people at the top (the experts who know how to use the tool) go higher, while the people at the bottom (those whose jobs the tool can do easily) get pushed down. Second, "informal employment" is like working in a backyard shed instead of a professional restaurant. These workers often don't have the same safety nets, like insurance or guaranteed pay, as the formal workers. This paper dives into the Indian kitchen to see if this new AI sous-chef is making the seesaw tilt dangerously, leaving some workers behind while others soar.
The Paper's Big Discovery: A Tale of Two Workforces
This research, titled "Long-Term Effects of Artificial Intelligence on Employment and Wages in India," tackles a tricky problem. The author, Kowser Ali Jan, wanted to look at real, private data from millions of Indian workers to see exactly how AI is changing their paychecks and job security. However, because that specific data is locked away in government vaults and couldn't be accessed for this study, the researcher had to get creative.
Instead of giving up, the paper uses a "dual-track" strategy. Track A is a highly realistic simulation. Think of it as building a giant, digital twin of India's workforce using a computer program. The researcher fed this program with the known facts about India's economy—how many people have degrees, how many work in cities versus villages, and what they usually earn—and then programmed the AI to act exactly as experts predict it would. Track B looks at the real, public numbers that are available, like the official unemployment rate and the number of new formal jobs being created.
Here is what the paper found, using these two tracks to tell the full story:
1. The "Skill Polarization" Seesaw is Tipping
The most important finding is that AI is not hurting everyone equally. It is creating a sharp divide based on education.
- For Graduates (The "Augmentation" Effect): In the simulation, workers with a university degree who use AI are seeing their wages go up. The data shows a massive triple-interaction coefficient of 0.907 (with a p-value less than .001). In plain English, this means that for graduates, AI acts like a super-power, boosting their value and pay significantly.
- For Non-Graduates (The "Displacement" Effect): For workers without a degree, the story is different. The simulation shows a wage penalty (a DiD coefficient of -0.126, p < .001). AI is doing the tasks they used to do, which lowers their pay or makes their jobs harder to keep.
2. The Danger of "Going Informal"
The paper also found that AI is increasing the likelihood of non-graduate workers ending up in the "backyard shed" of the economy—informal employment. A statistical test called a chi-square test showed a strong link between high AI exposure and informal jobs (χ²(2) = 234.90, p < .001). Specifically, the simulation indicates that the probability of informal employment rises for non-graduates working in high-exposure occupations. Essentially, as AI takes over certain tasks, workers without degrees are more likely to lose their formal, protected jobs and end up in unstable, informal work.
3. The "Aggregate" Illusion
One of the paper's most clever points is about the big picture. If you just look at the total number of jobs in India, it looks like everything is fine. The real data (Track B) shows that the unemployment rate dropped steadily from 6.0% in 2017-18 to 3.2% in 2023-24.
- However, the paper argues this drop happened before AI really took off. The trend was already going down, so you can't blame (or credit) AI for the total number of jobs yet.
- The real story is hidden in the details: The total number of jobs might be stable, but the quality of those jobs is splitting apart. The "average" job looks okay, but the gap between the high-paid, AI-augmented graduate and the low-paid, informal worker is getting wider.
4. What the Paper Rules Out
The paper is very careful to say what AI is not doing yet. It explicitly argues that the current drop in unemployment is not caused by AI. In fact, the real data shows that the unemployment rate was already falling fast before AI became a major force. The paper also clarifies that while AI is changing who gets paid and how much, it hasn't caused a massive, sudden wave of job destruction across the whole country yet. The changes are subtle, happening inside the "skill divide" rather than wiping out entire industries overnight.
The Bottom Line
This paper suggests that in India, AI is acting like a magnifying glass for inequality. It is making the educated, skilled workers richer and more valuable, while making it harder for those without degrees to keep their footing in the formal economy.
The researcher concludes that we shouldn't worry about AI destroying all jobs, but we should be very worried about the wedge it is driving between different types of workers. The solution, the paper hints, isn't just about stopping the technology, but about helping the workers who are being left behind—through better training, safety nets for informal workers, and policies that ensure the benefits of AI are shared more fairly. The simulation shows a future where the gap widens unless we actively work to close it.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.