Technology transformation

India should aim to gain from the change

The widespread adoption of generative artificial intelligence (AI) will lead to productivity gains along with big changes in employment patterns. A recent study by Goldman Sachs (GS) says about 18 per cent of full-time global employment — amounting to 300 million jobs — could be replaced by AI. The adoption could boost global productivity by 1.5 percentage points over the next 10 years. Over a decade, that would amount to gains of $7 trillion. While India would be among the nations where employment is relatively unaffected, it would also see low productivity gains. Productivity growth may be only 0.7 percentage points while 11-12 per cent of work in India could be automated. This could have worrying implications. Low productivity gains from a new technology may be disappointing while an 11-12 per cent reduction in employment opportunities will have severe implications, given a trend of widespread unemployment and underemployment.
 

Yet the current policy response seems pragmatic, though it may be underestimating the need for a policy push, given the likely political impact of widespread AI adoption. Union Electronics & Information Technology Minister Ashwini Vaishnaw recently informed Parliament the government was not currently planning to regulate growth or set laws for AI usage. He did, however, acknowledge ethical concerns and risks around AI and said the government had started making efforts to standardise responsible AI and promote best practices. The GS study is focused on generative AI — such as ChatGPT — which has generalised use cases and expands the range of tasks. Given the ability to generate human-seeming content output across speech, audio, and video, and a comfort with natural language instructions, generative AI is easy to use. In addition, the study also referred to exponential increases in raw computing power, which enables rapid gains in the complexity of tasks AI can perform.
 

A large component of clerical work across government departments and courts can be automated with efficiency gains but loss of permanent employment. The algorithms of generative AI have already surpassed the human average for tasks like image classification, and even reading comprehension. The study concluded AI automation would have the lowest impact in India, followed by Kenya, Vietnam, Nigeria, Mainland China, and Thailand. In these nations, the work that could be automated ranges from 11 per cent to 16 per cent. The highest impact will be in Hong Kong, followed by Israel, Japan, Sweden, the US, and the UK. In these nations, the percentage of jobs that could be automated ranges between 25 and 30 per cent. In general, developed nations are far more exposed with two-thirds of current jobs exposed to some degree of automation.
 

Detailed analysis of the US and Europe indicates 46 per cent of office and administrative functions, 44 per cent of legal work, and 37 per cent of architectural and engineering work can be automated. The last is particularly interesting since these are skilled, “creative” tasks where complicated mathematical calculations, as well as visuospatial acuity, are required. Not imposing regulatory constraints could help India develop AI-based capacities. However, given the inevitability of AI-induction, policy needs to be more pro-active. By encouraging India’s large IT workforce to focus on AI research, policymakers should look to accelerate productive gains, and ensure alternative employment opportunities. Given that large corporations are establishing capability centres in India, it has the opportunity to lead and gain from the evolving technology transformation.