AI dissonance: Politics over artificial intelligence is a phenomenon that can’t be wished away

This trend, apart from the polarizing debate over an AI bubble bursting or not bursting, is likely to dominate the AI policy landscape. This has consequences for India too, which has been judged as an ‘AI loser’ by sections of global markets.

An AI bubble is a top-of-the-mind issue today. However, it is perhaps premature to debate the likelihood of a bubble or non-bubble at this juncture, let alone its bursting. A lot will depend on the nature and speed of AI adoption and its manifestation in enterprise bottom-lines.

Much money has already been sunk into AI infrastructure, products and services, and a lot more has been committed for the next three-four years. Consequently, most financial institutions seem reluctant to accept the eventuality of a meltdown.

As a precautionary note, though, it might be wise to recall 2007, when most traders were in denial about the likelihood of a market meltdown. History remains the best starting point for building a risk-mitigation model.

Bubble or not, one thing is certain: AI is here to stay and governments across the globe will have their hands full trying to erect meaningful guard-rails around this emerging technology. These efforts are bound to collide with industry resistance. This new flashpoint is likely to define 2026.

Battlelines have already been drawn. Silicon Valley’s AI tech-preneurs are amassing a fund to sponsor politicians promising light-touch AI regulation.

A political action committee—Leading The Future, backed by venture capital firm Andreessen Horowitz and OpenAI co-founder Greg Brockman, among others—has raised $100 million to finance the election campaigns of industry-friendly politicians during the November 2026 mid-term elections in the US.

The pro-regulation section is not sitting on its hands. Two former Congressmen—a Republican and a Democrat—have launched Public First, with an initial $50-million kitty to identify pro-regulation candidates on both sides of the aisle.

Individual states in the US are headed for a showdown with the federal government. A new California bill requires AI companies to publish and abide by safety policies for managing AI risks.

A New York bill sought to improve on this bill, but industry influence managed to water down the text of the final legislation.

A White House executive order has further muddied regulatory waters by trying to rein in states’ power to regulate the AI industry, but that comes after almost every US state has already passed AI regulation, with over 100 laws legislated between them all. Expect feathers to fly.

India, on the other hand, has started its AI regulatory journey on a circumspect but predictable note: policies are tilted in favour of industry rather than citizens, as they focus more on governance than regulation.

The government’s recently released AI Governance Guidelines rely on existing laws, such as the Digital Personal Data Protection Act, instead of enacting separate AI legislation.

While the government has committed to intervene if AI models cause harm, there is apprehension that the absence of a stand-alone law will leave room for interpretation and misuse.

The Indian government’s soft-touch regulatory regime may have been necessitated by market compulsions and strategic impulses. The AI Governance Guidelines have “innovation” as their centrepiece and wish to be viewed more as an enabling provision than a regulatory decree.

The motivation behind this policy mix seems to be driven by a global perception of India as an ‘AI loser,’ a view that may have led to a record retrenchment this year by foreign portfolio investors in favour of AI ‘winner’ markets such as South Korea and China. The new guidelines try to reverse that sentiment.

There is some realization that India may have missed critical layers of the AI stack. The AI industry has three broad layers: infrastructure (chips, servers, data centres, energy grids), software (large language model) and, finally, products and services for enterprise clients.

India has seen investments come into the first layer, with Google, Microsoft, Amazon and Meta committing large sums to AI data centres. But India lacks large homegrown AI models, like US-based OpenAI’s or China’s DeepSeek, which require deep and patient investments.

The government has committed ₹10,300 crore ($1.15 billion) for supplying subsidized chips to startups or research centres for developing AI platforms that exclude the Western biases of existing models.

This puny outlay, compared with other sovereign budgets (such as France’s $117 billion), is predicated on the hope of private capital pitching in.

That might be slightly problematic, given India Inc’s reluctance to invest; a survey by EY India and Confederation of Indian Industry shows that over 95% of enterprises have earmarked less than 20% of their infotech budgets for AI.

But the government’s year-end guidelines reveal an indulgent attitude towards the private sector in the name of innovation that may tempt many Indian companies to use artificial intelligence for shirking their after-sales service responsibilities.

Many banks and consumer-facing companies have already replaced a human voice with under-developed, inefficient and untrained AI bots. This promises to stir up a different kind of politics over AI.