SEBI’s planned AI/ML framework will require capital-market firms to document model governance, deploy kill switches and keep humans in the loop, aligning with IOSCO standards. It aims to curb flash-crash risk from runaway algorithms, though smaller firms may face higher compliance costs and derivative settlement changes loom.
What Happened
On October 10, 2026, the Securities and Exchange Board of India (SEBI) announced that it will soon publish guidelines governing the use of artificial intelligence (AI) and machine learning (ML) in capital markets. Chairman Tuhin Kanta Pandey revealed that the framework will adopt a tiered approach, featuring clear accountability, robust data controls, kill switches and explicit human oversight. He added that the guidelines will align with the International Organization of Securities Commissions’ (IOSCO) supervisory toolkit for AI, aiming to support an agile governance structure.
In addition, SEBI will issue rules on the settlement price of derivatives on expiry days within about a week, following a review of comments received on a consultation paper.
“Technology cannot dilute accountability. Regulated entities remain responsible for AI outcomes, data quality, model governance and cyber resilience,” Pandey said, emphasizing that human oversight must stay front‑and‑center as technology reshapes market operations and supervision.
He also highlighted the inclusion of kill switches in the proposed framework to allow rapid shutdown of malfunctioning AI systems.
What This Means For You
For market participants, the upcoming guidelines will set the legal and operational baseline for deploying AI in trading, risk management, and compliance. If you run a brokerage, you must review your AI models for compliance with the new tiered accountability structure. This means documenting model development, validating data integrity, and establishing clear escalation paths for model failures.
Compliance teams should prepare to implement kill switches and human‑in‑the‑loop checks. A kill switch is a mechanism that can instantly halt an AI system if it behaves unexpectedly. Integrating such controls will likely require updates to your IT architecture and incident‑response playbooks.
For algorithmic traders, the guidelines will influence how you design and test predictive models. You’ll need to prove that your models meet IOSCO’s supervisory standards, which may involve third‑party audits or certifications. This could increase upfront costs but will also reduce the risk of regulatory penalties.
The derivative settlement rule update is particularly relevant for firms handling options and futures. Expect new documentation on how settlement prices are calculated on expiry days, which could affect hedging strategies and end‑of‑day reconciliation processes. Prepare to adjust your pricing engines and audit trails to capture the new requirements.
Financial institutions should also note that the guidelines will cover cyber resilience. This means reinforcing data protection protocols, ensuring secure data pipelines, and conducting regular penetration tests focused on AI components.
In short, you should:
1. Map your AI assets to the proposed tiers and identify accountability gaps.
2. Deploy kill switches and human‑override mechanisms.
3. Align data governance practices with IOSCO’s toolkit.
4. Update derivative settlement processes to comply with the new rules.
5. Strengthen cyber resilience measures around AI systems.
Why It Matters
The guidelines signal a shift toward stricter oversight of AI in financial markets, reflecting global concerns about algorithmic transparency and systemic risk. By embedding kill switches and human oversight, SEBI aims to prevent runaway model behavior that could destabilize markets. This approach echoes the recent AI in Cloud‑Native Banking Platforms in Banking piece, where the focus was on ensuring robust governance in AI‑driven banking services.
Moreover, the alignment with IOSCO’s toolkit positions India as a leader in harmonizing AI regulation across jurisdictions. This could attract international capital flows from firms seeking a clear regulatory environment. On the other hand, the new rules may increase compliance costs, especially for smaller firms that rely heavily on proprietary AI models.
From a systemic perspective, the guidelines could reduce the probability of flash crashes triggered by faulty AI models. By mandating rigorous testing and human oversight, SEBI is proactively addressing the risk that AI systems can amplify market volatility.
In the broader regulatory landscape, this move complements global efforts to standardize AI governance, such as the EU’s AI Act and the U.S. Federal Reserve’s AI risk framework. It also dovetails with the AI Verification Tax: Why Efficiency Is Falling Now article, which highlighted the costs of inadequate AI oversight.
Key Takeaway
- SEBI will publish AI/ML guidelines with a tiered accountability model, kill switches, and human oversight.
- New derivative settlement rules are expected within a week, affecting end‑of‑day pricing.
- Compliance teams must audit data quality, model governance, and cyber resilience to meet IOSCO standards.
- Implementing kill switches and human‑in‑the‑loop controls will become mandatory for AI systems in trading and risk management.
Leave a Reply