AI and the Future of Employment
There is heated debate on the future of artificial intelligence in the West. India must make its voice heard
Published in the Indian Express
Over the weekend, a friend who works in the tech industry in Silicon Valley gave me a demo of the latest AI language model, GPT-4. I had played around with its predecessor, GPT-3 but I was blown away by the advanced capabilities of the new model. According to OpenAI, GPT-4 exhibits "human-level performance" on various professional and academic benchmarks, passing a simulated bar exam with a score in the top 10% of test takers. What makes this specially mind-blowing is the pace of change: just months back, the older version ChatGPT-3 cleared the same exam with a score in the bottom 10% of all scores. As my friend showed me how GPT-4 could automate tasks like writing code, testing, debugging, drafting emails and notes, I felt a sense of alarm. It became clear to me that almost any job which doesn't require physical presence, could soon be substantially automated and done through generative AI. In fact, a recent report by Goldman Sachs estimates that as many as 300 million full-time jobs around the world could be automated in some way by the newest wave of artificial intelligence.
This has troubling implications for India. 65% of the population is below the age of 35 and India is already struggling with unemployment. It has been widely reported that the country must create at least 10 million jobs annually until 2030 to reduce its rate of unemployment and underemployment. However, this job creation goal which was ambitious to begin with is now further at risk: entire categories of jobs that India would have hoped to create are now in danger of being lost due to automation. These include offshored jobs like entry-level programming, customer service, documentation, and more, which gave India’s tech industry its initial headstart and made India the world’s largest back office.
It's not clear whether our decision-makers in India - whether it is our political class or industry leaders - are fully cognizant of the implications of these developments and thinking of how best to respond to them. In the tech industry, there can be an airy casualness around "innovation," as if all tech industry developments will necessarily lead to societal progress because they increase "productivity." While it's true that these developments will increase productivity of the smaller number of individuals who will survive the coming waves of automation, it is likely that this will also lead to jobs getting eliminated and people becoming unemployed. Past experience tells us that individuals who become unemployed due to automation aren't necessarily absorbed into other more productive jobs. In fact, they usually just stay unemployed. This is due to both our inability to retrain individuals at scale for higher order jobs and the lack of enough higher-order jobs to absorb everyone who has been rendered unemployed or unemployable in the first place. On the political front, the discourse around unemployment is also limited. In election after election, there are rhetorical announcements of lakhs of new jobs in the future and the promise of a paltry unemployment allowance for India's youth. However this framework barely scratches the surface: the job creation numbers lack credibility in the face of mounting job losses; the amount of unemployment allowance is pitifully meager at an individual level but even then unsustainable at the level of state finances, lacks the dignity of gainful employment, and leaves our youth restless.
There has been some discussion on the need to put guardrails around AI to limit its harms. However, these conversations often focus on the impact of AI on disinformation and sci-fi scenarios of AI developing "independent goals" which work against human interest. While these conversations are certainly important, the possibility of AI leading to large-scale unemployment is no longer just a theoretical construct in the future. The last decade has shown that the pace of development in the tech industry has mostly outpaced the regulatory ability of policymakers globally and we need to initiate an urgent national debate on this issue if we are to manage this transition in a smooth manner. This is especially urgent because on this particular issue, India does not have a lot of regulatory room - these developments are happening outside its shores (mostly in the United States and China) and a significant chunk of the impact at first will be in jobs which do not get offshored to India. It is thus important to think through the levers we have at our disposal and the social infrastructure we need to put in place today to manage what many are calling - without being hyperbolic - one of the biggest disruptions the world has seen. It is thus important to expand the national discussion beyond the usual rhetoric of innovation and productivity to larger questions of the impact of AI on employment and the role of government and industry in mitigating its effects. This includes fundamentally rearchitecting how we educate and skill our youth population, building supportive infrastructure, protecting workers’ rights, and expanding the kind and nature of work we value. Finally, these developments and their impact will be global in nature. There is already heated debate on the future of artificial intelligence in the developed world, India must make its voice heard.
Also Read: AI Search Chatbots May Transform The Internet As We Know It. Not Necessarily For Good.