Most Heated Indian AI Debates of 2025

2025 was supposed to be the year AI went truly mainstream in India. Instead, it turned into a spectacle of overhyped products, hollow promises and swift public backlash, which led to debates.

The discussions came thick and fast, revolving around failed business promises, misleading marketing pitches, culturally tone-deaf moves, biased algorithms, and AI-generated content that veered off course. Each stumble was amplified in real time on social media, turning hype into ridicule.

Here is a look at the most talked-about controversies around AI this year, debated and dissected endlessly on X and LinkedIn.

Sarvam AI’s Rocky Launch

India’s billion-dollar AI startup Sarvam AI launched its flagship multilingual model, Sarvam-M, in 2025 with the aim of supporting Indian languages. The company clarified it was not the final product, with the full foundational model expected early next year. Sarvam-M, built on Mistral, disappointed early users and managed only 23 downloads in its first two days.

Comparisons to a Korean college project that got two lakh downloads per month further made the contrast starker. Critics called it “embarrassing” and demanded more from the company. Since then, the company has worked on rebuilding trust and has recorded more traction and weekly downloads on Hugging Face.

Builder.ai’s Fake AI

Builder.ai entered 2025 claiming to revolutionise AI-powered app development, but was exposed as a massive fraud. Investigations revealed the company was essentially employing 700 engineers to write code manually despite being marketed as AI-driven. On top of that, revenues were allegedly inflated by 300%, leading to bankruptcy.

Following this, customers were locked out of their own apps and source code. The fallout is now regarded as the biggest AI startup collapse since the ChatGPT boom began, sparking discussions around inflated valuations and why global investors often hesitate to back Indian startups.

Raanjhana’s AI-Generated ‘Happy Ending’

Raanjhanaa, the 2013 romantic tragedy that etched itself into public memory with its heartbreaking ending, made a surprising return to theatres under its Tamil-language title Ambikapathy. But this time, the audiences were met with a very different conclusion. Eros Media Group re-released the film with a machine-generated happy ending, created using AI.

The move quickly drew criticism on social media. Film director Aanand L Rai publicly disowned the altered version, calling it unauthorised and distancing himself and his team from the project.

“To watch Raanjhanaa, a film born out of care, conflict, collaboration and creative risk, be altered, repackaged and re-released without my knowledge or consent has been nothing short of devastating,” Rai wrote on Instagram. “What makes it worse is the complete ease and casualness with which it’s been done.”

Puch AI’s $50 Billion Counter-Offer

Puch AI, founded by Siddharth Bhatia, who is also the founder of TurboML, promised a sovereign AI platform tailored to Indian languages and cultural nuances. The company drew attention early on by buying India’s most expensive phone number (9090909090) and introducing celebrity voice features.

However, users soon discovered that the voices were inconsistent and often failed to respond meaningfully. Meanwhile, several followers praised Puch AI’s attempt at creating an AI model accessible on WhatsApp.

Controversy deepened when the company made a headline-grabbing $50 billion counter-offer to Perplexity AI, sparking debates over its legitimacy and business practices. Moreover, Puch AI’s also allegedly clashed with Reliance Jio over trademark violation and redirection of a domain to the Puch AI platform.

Founder Siddharth Bhatia later told AIM that Jio had blocked Puch AI on its network, a claim that quickly spread across social media.

Grok Confronts Racist Remarks

In 2025, Grok, Elon Musk’s AI chatbot, found itself in the midst of a debate in India after responding to racist comments targeting Indians. The controversy began when an X user asked Grok about which demographic in America posed the “biggest problem”, followed by another query on which group was the most productive.

Grok responded with data, showing that Asian Americans are among the most productive communities, with the highest median weekly earnings of $1,474, the lowest unemployment rate at 3% and significant contributions to innovation and the economy.

The exchange escalated when a user singled out Indians, labelling them as “cancer”. Grok countered with facts, highlighting that Indians are the top earners among Asian Americans, with a median household income above $150,000, and that they drive innovation in tech and business. The AI called out the user’s bias, even responding wittily to further provocations with, “Just circuits and code spitting facts.”

DeepSeek’s Reluctance to Talk About Arunachal

In 2025, Chinese-developed AI chatbot, DeepSeek, faced criticism in India for its reluctance to provide information on Arunachal Pradesh and its indigenous Tani tribes. Users reported that the AI repeatedly evaded questions about the region and tribes, giving vague responses such as, “Sorry, that’s beyond my current scope. Let’s talk about something else.”

The omissions went further. DeepSeek left Arunachal Pradesh out when listing Northeast Indian states and avoided sensitive geopolitical topics, including the 1962 Sino-Indian war and Chinese territorial claims over Arunachal and Aksai Chin.

Experts warned that DeepSeek’s behaviour could undermine India’s sovereignty and misrepresent the cultural and demographic identity of the Tani tribes, raising difficult questions about bias and geopolitics embedded in AI.

Dhruv Rathee’s AI Fiesta

India’s much-hyped AI ‘super-app’ Fiesta launched with promises of access to every major model for ₹999 a month. In reality, users were limited to a small number of tokens, experienced sluggish performance and impractical integrations. OpenAI’s cheaper ChatGPT Go launched the next day and wiped it out completely.

Fiesta’s gimmicky buzz lasted barely 48 hours. Several Indian developers dismissed it as unhelpful and noted it could not be called a fully Indian product. The platform faced criticism for its subscription pricing, token system and misleading marketing.

“Apparently, a much bigger YouTube creator dropped their version of T3 Chat? Good luck to them. This shit isn’t easy,” Theo Brown, YouTuber and CEO of T3 Chat, said. Meanwhile, others like Varun Mayya, responded by offering their congratulations.

Users accused Rathee of overpromising features such as access to GPT-5, while only delivering GPT-4, leading to debates about the platform’s credibility.

The post Most Heated Indian AI Debates of 2025 appeared first on Analytics India Magazine.