What happened
The Securities and Exchange Board of India (SEBI) is preparing a set of guidelines aimed at regulating the use of artificial‑intelligence‑powered trading agents in Indian capital markets. The draft framework focuses on transparency, accountability, human oversight, risk‑management controls and mechanisms to preserve market stability. It builds on a June 2025 consultation paper that outlined broad principles for AI/ML use in securities markets, and seeks to address emerging concerns such as autonomous strategy changes, the difficulty of supervising adaptive systems, and the potential for multiple AI agents to amplify market moves during periods of stress.
SEBI announced that it is drafting a regulatory framework specifically for AI‑driven trading agents. The draft emphasizes four pillars: accountability (identifying who is responsible for AI‑generated trades), transparency (requiring disclosures about AI usage and decision‑making processes), human oversight (mandating mechanisms for manual intervention when needed), and risk management (including safeguards against systemic risks).
The regulator highlighted that AI agents differ from traditional algorithmic trading because they can autonomously adapt to new data, potentially altering their behaviour after deployment. This raises challenges for supervision, approval processes, and attribution of liability for trades executed by such systems.
A key concern outlined by SEBI is the possibility of multiple AI agents reacting to the same market signals simultaneously, which could amplify buying or selling pressure during periods of stress. The draft proposes monitoring and reporting requirements to detect such clustering effects.
The proposed guidelines build on a June 2025 consultation paper that already covered data security, independent audits, and disclosure obligations for AI/ML applications in securities markets. The new draft seeks to translate those principles into concrete rules for trading agents.
SEBI clarified that the draft is not yet final and that the ultimate regulatory requirements will be set out in the formal guidelines once the consultation process concludes.
Source details: insights.dsij.in
Why it matters
AI‑enabled trading agents can analyse vast data sets and modify their strategies in real time, a capability that differs from traditional rule‑based algorithms. While this flexibility can improve efficiency, it also creates regulatory blind spots: the system’s behaviour may evolve after deployment, making it harder for regulators to trace decisions, assign responsibility, or intervene when harmful patterns emerge. SEBI’s proposed rules aim to close these gaps by mandating clear disclosures about AI usage, record‑keeping that enables reconstruction of automated decisions, and mandatory human‑in‑the‑loop controls for high‑impact actions. By establishing such safeguards, the regulator hopes to prevent unintended market disruptions—such as synchronized selling by multiple AI agents during volatility spikes—and to protect retail investors from opaque, potentially manipulative trading practices. The guidelines also signal to Indian financial firms that innovation must be balanced with robust governance, potentially shaping how AI is adopted across trading, compliance and surveillance functions.
AI‑driven trading agents can process far more information than human traders or static algorithms, potentially leading to faster and more efficient market responses. However, their adaptive nature also means that they can develop unforeseen behaviours, making it difficult for regulators to monitor compliance and for investors to understand the risks they face.
By requiring detailed disclosures and record‑keeping, SEBI aims to create an audit trail that can be examined after market events, helping regulators reconstruct the decision‑making chain and assign responsibility. This is crucial for maintaining investor confidence and ensuring that market manipulation or systemic shocks can be traced and addressed.
Human‑in‑the‑loop provisions are intended to prevent runaway automation. In practice, this could mean that firms must implement kill‑switches or manual approval steps for trades that exceed certain risk thresholds, thereby adding a layer of protection against unintended market impacts.
The guidelines could set a precedent for other emerging markets grappling with similar AI adoption challenges. If effective, SEBI’s approach may influence regional regulatory bodies and contribute to a more harmonised global framework for AI in finance.
Finally, the focus on systemic risk—specifically the risk of multiple AI agents acting in concert—addresses a novel market‑wide vulnerability that traditional regulation does not cover. Managing this risk is essential for preserving market stability, especially as AI adoption accelerates.
Interactive Mechanism: How It Actually Works
Explore the underlying technology behind this development interactively.
crm_get_transaction(id='4092').Why can ethical evaluation not be reduced to one model score?
What to watch next
Stakeholders will watch for the final issuance of SEBI’s guidelines, the timeline for compliance, and any enforcement mechanisms that accompany them. Market participants may also monitor how Indian broker‑dealers and asset managers adjust their AI systems to meet the new transparency and oversight requirements, and whether similar regulatory approaches emerge in other jurisdictions. Additionally, the impact on market liquidity and volatility during the transition period will be a key metric for assessing the effectiveness of the proposed safeguards.
The timeline for SEBI’s final guidelines, including any public consultation periods and the expected date of enforcement.
How Indian brokerage firms, mutual funds, and hedge funds will adapt their AI trading platforms to meet the new transparency and oversight standards.
Potential pushback or legal challenges from industry groups concerned about the cost and operational impact of compliance.
Whether SEBI will introduce specific penalties for non‑compliance, and how those penalties compare to existing rules for algorithmic trading violations.
The emergence of similar regulatory initiatives in other major markets (e.g., the United States, European Union, China) and any coordination efforts that may arise.