Alluring Promise Of AI To Raise Productivity Rather Than Kill Jobs

The week in review has been notable for decline — of stock market indices, the rupee, and economic growth. In Canada, Justin Trudeau, for long the Prince Charming of the world’s liberals, also declined, in charisma and power. He resigned as leader of the Liberal Party and prime minister. This is likely to lead to elections earlier than the ones scheduled for November.

In the US, president-elect Donald Trump continued to win friends and influence people, threatening to take over the Panama Canal, annex Greenland, and incorporate Canada as the 51st of the states that together constitute the US of A. Panamanians, Canadians, and Danes bridled. British premier Keir Starmer broke form to speak with passion, rather than with platitudes, in response to Trump’s presumed alter ego Elon Musk’s criticism of Starmer, for allegedly failing to prosecute a ring of child abusers during his term as director of public prosecutions.

China, reports the Financial Times, is moving to monetary policy that is similar to what market economies use, relying on a policy rate and shedding quantitative credit targets for specific sectors.

In India, the government notified draft rules to enforce the Digital Personal Data Protection Act. The law and the rules give the state unfettered access to personal data in the name of national security, defence of sovereignty, and friendly relations with other countries. It seeks to shift the responsibility for the potential harm that some corners of social media might cause children from social media companies to the parents, whose consent is now needed and who would be required to part with their particulars in order to establish that they are, indeed, the parents of the children in question.

In a positive development, the government tweaked some rules of the National Pension System, allowing subscribers to stay invested till they reach age 75 and withdraw in staggered instalments the portion of their saved corpus that is not used for mandated purchase of an annuity.

Growth Falters

The first advance estimate for economic growth in the current fiscal year came in at 6.4%. After growth for the June-September quarter had been estimated at 5.4%, different agencies have, indeed, been revising downward the growth rate for the entire year. India’s central bank, the Reserve Bank of India, brought down its growth estimate from an optimistic 7.2% to 6.6%. But the official estimate by the National Statistics Office has been lower still. It estimates that only agriculture and government expenditure are likely to register faster growth than in the past year.

The government itself is culpable, for the most part, for the slowdown in growth. While it had made much, in July this year, of its big budget outlay on capital expenditure, the actual spending has been relatively anaemic. While, as of November, the government had realised 59% of its estimated revenue for the fiscal year, it had managed to spend only 46% of its budgeted outlay on capital expenditure. The estimate for the whole year expects gross fixed capital formation to barely cross 30% of GDP (at current prices). During the period of fast growth under the previous regime, this measure had stayed in the mid-thirties year after year. China, during its sustained period of double-digit growth, had registered investment rates close to or over 50% of GDP.

As this column has argued in the past, the government has to reintroduce a robust policy of public-private partnership in building infrastructure. Infrastructure is the only place where the private sector has any sane reason to invest at the moment, considering that capacity utilization has been stuck at or below 75% for nearly a decade, and given the recent slowdown in sales. That must be twinned with an overhaul of the debt market, unifying the government bond and corporate bond segments under a common regulator, Sebi.

The Debt Market Imperative

The debt market, rather than bank loans, must finance infrastructure – for two reasons. Infrastructure projects have long gestation periods. If banks finance such projects out of their relatively shorter-term liabilities, that is, deposits, there would be considerable mismatch between the temporal profiles of bank liabilities (relatively short) and bank assets, the loans extended to infrastructure (relatively long).

The other reason why it is preferable to fund infrastructure projects with debt raised via bonds, rather than with bank loans, is the transparency and accountability it would enforce on such financing. Multiple analysts, mutual fund and pension fund managers would evaluate the finances of the projects raising debt. In the case of bank loans, a loan sanctioning committee is all that needs to be satisfied, about the soundness of the project costing. It is possible to deploy extraneous considerations to ‘persuade’ a compact committee, while it is more difficult to sway a large distribution of independent analysts and fund managers.

AI To The Rescue?

Microsoft boss Satya Nadella visited India to announce a slew of investments in developing artificial intelligence (AI) and AI capability, totalling $3 billion. The maker of chips powering AI, Nvidia, also announced a major expansion of its operations in India.

Stock markets around the world have tended to come out of the spell that AI had cast on investors, causing them to loosen their purse strings and buy into the AI story as they understood it. Assorted sharp practitioners have profited from the herd preference for AI. Now that it is becoming increasingly clear that not every company that adds the appellation AI to itself is going to turn in magical improvements in productivity and profits, investors are turning more discerning about how they chase their AI dreams.

Global surveys on the use of AI — these are many and often show inconsistent results — suggest that right now, individual employees are more likely to make use of AI tools to get their work done faster and better, than for companies to make concerted investments in AI. This could change in the coming days.

Microsoft seeks to persuade its customers to adopt its Copilot suite to create a variety of sector- and use-specific Copilot agents, to automate a lot of grunt work and generate actionable output. The software maker already has the cloud infrastructure needed to host the massive data processing AI calls for, and promises to bundle in transparency, security, and energy efficiency with their AI offering.

Rival AI providers can be expected to make similar offerings to customers as well. Just as most people make extensive use of their smartphones — supercomputers far more sophisticated, each, than the computers that NASA used to land a man on the moon — without a clue as to how these devices work, they would be able to make use of AI tools even if they have no clue as to how these work.

Nadella is sanguine about the trade-off AI will pose on job generation and job destruction. All that will happen, he says, is that AI will raise both the threshold and the peak of the typical distribution of employee productivity. Of course, this does not say anything about those who insist that the last thing they will learn in life would be what they learnt for their final qualifying exam at university.