Emphasis should be to build strong AI guardrails, not just compliance checkboxes: Grant Thornton Bharat

As artificial intelligence transitions from a frontier for experimentation to becoming part of the regulated infrastructure, the emphasis should be on building strong guardrails that are not just compliance checkboxes, but are enablers of trust, transparency, and long-term value creation, Grant Thornton Bharat said in a newly released report.

“AI is no longer an unexplored frontier; it has the potential to reshape how financial services are designed, delivered, and governed. The recent frameworks released by the RBI and Sebi mark a decisive shift in India’s regulatory approach, from observing innovation to orchestrating it responsibly,” reads the report.

One of the key findings of the report is that AI has evolved from an experimental tool to a regulated infrastructure, which necessitates fairness, transparency, and effective governance. Regulators in India are also shifting from observation to orchestration, promoting innovation with strong guardrails for accountability and consumer protection.

The report contextualises the inception of artificial intelligence since the 1960s, from which it has experienced alternating cycles of intense innovation followed by periods of stagnation. The 1990s and 2000s marked a significant shift from rule-based systems to statistical methods and machine learning (ML), laying the groundwork for the breakthroughs of the 2010s, when deep learning revolutionised fields like speech recognition and autonomous vehicles.

In the 2020s, it took a further step forward with the integration of massive computing power and sophisticated algorithms, ushering in the era of Generative AI and unlocking unprecedented capabilities. However, Grant Thornton cites an RBI report that says that only about 20% of regulated financial entities in India have adopted AI technologies, primarily limited to basic rule-based or moderately complex ML models. “This cautious adoption is attributed to various challenges such as a shortage of skilled AI talent, high implementation costs, limited access to quality data and computing infrastructure, and regulatory uncertainty.”

Financial institutions face four key structural challenges: data quality issues, infrastructure gaps, talent shortages, and regulatory ambiguity. The report outlines a compliance roadmap for banks, NBFCs, insurers, fintechs, and SROs to responsibly integrate AI into their business models.

For banks and NBFCs, AI becomes a board-level governance issue, demanding model risk management and fairness programs. For smaller institutions, national AI infrastructure, such as AI Kosh and IndiaAI Compute Platform, will bridge capacity gaps. For capital markets, transparency and auditability will be key to establishing trust in AI-led decision-making. For insurers and fintechs, innovation must happen under supervision — with fairness, explainability, and consumer protection at the core.