Airtel Is Giving Free Perplexity AI. Can India Handle The Pandora's Box It'll Open?

Bharti Airtel was announcing a year-long free subscription to Perplexity Pro, an artificial intelligence (AI) powered search and answer platform. Out of pocket, it would cost about Rs 17,000. Not a cheap giveaway. 

This has become the first instance of an Indian telco bundling generative AI at this scale. The partnership gives over 390 million users the chance to access a premium AI tool. 

That also gives 390 million people a reason to ask tough questions about AI regulation, data collection, and privacy. But, in the face of a free magic lamp, few feel the need to. 

Bundle Of Opportunity Or Dependence?

The model of bundling subscriptions with telecom services isn’t a new phenomenon. Think of how telcos bundled OTT platforms, giving access to platforms like Netflix, Disney+ Hotstar, Amazon Prime Video, and SonyLIV.

It’s a proven strategy to lure users in a market where only about 18 per cent of India’s 547 million OTT viewers pay for subscriptions.

For Perplexity AI, which lacks a device ecosystem of its own, this Airtel deal is a gateway to mass adoption. 

Mohit Chawdhry, technology policy and strategy professional, told The Secretariat why this deal was conjured. "For Perplexity, it's the easiest way to get on millions of devices in India,” he said. 

Globally, similar moves have already emerged. SoftBank in Japan and SK Telecom in South Korea have announced limited-time partnerships with the AI startup to expand access. In Germany, Deutsche Telekom is working on an “AI phone” with Perplexity integration apart from the free pro subscription rollout. 

In India, Reliance Jio is reportedly in talks with OpenAI to bring ChatGPT access to its users. But Airtel’s offer is the first in India and, for now, the most sweeping to make generative AI a go-to solution in everyday life. 

Access Or Ask Less?

India’s 806 million internet users and its growing Tier-2 and Tier-3 cities present a vast new frontier for AI adoption. Chawdhry sees this as a net positive step. “If you're a user who previously hadn’t had access to any of these top models, you now have it,” he said.

This mass exposure could spark digital literacy, new forms of learning, and productivity, especially where language models offer localised support. 

"It can help more people in India access information and use tools for things like translation, research, content creation, etc. This is particularly useful in a country like India, where mobile phones and affordable internet have reached a large part of the population, but many still lack access to advanced digital resources," Kalindhi Bhatia, AI regulation expert and partner at law firm BTG Advaya tells The Secretariat. 

But, as the unattributed popular maxim goes, "If you're not paying for the product, you are the product." 

"The downside is that it could lead to more digital tracking and give a few big companies too much control over what information people see and use," she says. 

Critics worry the model is less about equitable access and more about creating product dependency, which is getting users “hooked” before charging them, once the free period ends. Because it will end. 

Opt Out 

Perplexity, like most AI platforms, collects user interactions to improve its models. While users can toggle off data sharing, Chawdhry warns: “The average user is probably not going to do that.”

How often have you read the full privacy policy of something you’ve signed up for? Probably the same number of times as Indian telcos have considered pausing a rollout over privacy concerns.

"Unlike the EU and some US states, there is currently no central legislation specifically governing the use of AI or its integration into telecom infrastructure," says Bhatia.

"Consequently, any new AI based product is typically evaluated through the lens of existing laws relating to data privacy, intellectual property, contract law, and criminal liability. However, these frameworks were not designed with modern technologies like AI in mind, and they often fall short in addressing the unique challenges posed by these innovations," she says. 

India’s Digital Personal Data Protection (DPDP) Act was passed in 2023 to curb such ambiguity. But, the law has not yet been operationalised. Until it is, enforcement depends on the outdated IT Act which lack a rights-based framework.

"This means there are still very few regulatory requirements around how AI services use personal data" explains Bhatia. 

This raises concerns about what happens when newer users unfamiliar with AI risks engage with the tool. “There is always the red flag,” Chawdhry notes, “if users turn to it for health advice or personal guidance.”

Both Airtel and Perplexity’s terms permit the use of user data for service improvement, and potentially for commercial purposes. 

"If one is using a bundled AI tool offered by a telecom provider, it is important to stay alert," warns Bhatia. "A good practice is to follow ‘data minimisation’. Only share the bare minimum needed for a task and avoid entering personal, financial, or health-related information," she says. 

While Airtel clarifies that it “does not sell personal information”, it may share data with affiliates and partners. Whether that includes behavioural or AI prompt data is unclear.

Policy Reckoning On The Horizon

India’s AI-telco convergence is likely just beginning. Jio and Vodafone-Idea may follow Airtel’s lead. Perplexity claims to be working with over 25 telcos globally, suggesting this model may become the norm in emerging markets.

This also means regulators will need to play catch-up. India’s net neutrality framework prevents preferential data rates for content (as in the Free Basics case), but does not address bundling of services like AI behind general mobile plans. Nor does it address meaningful consent, a concept the DPDP Act claims to champion but has yet to enforce.

“It’s neither a revolution nor a trap — just marketing,” says Chawdhry. But it’s also a stress test for India’s data protection ambitions and a signal that AI regulation can’t afford to wait. 

At the same time, the real test will be whether millions become smarter AI users or simply generate more data for tech giants while losing their own brain cells. 

We can only hope that the seeming magic lamp with 300+ daily wishes (read: prompts) doesn’t turn out to be a pandora’s box.