Introduction
The increasing use of Artificial Intelligence in securities markets has transformed how trades are executed, but this shift has also led to the rise of automated market manipulation. This technological change is highly relevant to Indian capital markets, where automated systems now drive the majority of trading volume, highlighting why SEBI’s Prohibition of Fraudulent and Unfair Trade Practices (PFUTP) Regulations may be inadequate.
This article addresses a vital research question: Can SEBI’s existing PFUTP framework effectively regulate AI-driven securities manipulation? It tests the hypothesis that the PFUTP framework is insufficient because its dependence on proving human intent is disrupted by autonomous software. Structurally, the article first examines the rise of automated abuse, then analyses the legal limits of the PFUTP, and finally proposes essential regulatory reforms.
The Rise of Automated Trading and Market Abuse
As the automated trading market in India is growing rapidly, we must assess its extent to see why our current laws are struggling. In 2025, its total value was 615.61 million USD, and it is expected to grow to 1,350.34 million USD by 2034 at a 9.12% CAGR. This massive expansion has indeed altered the way stock exchanges function. For example, on the National Stock Exchange (NSE), automated trades have, for the first time in 2024, dominated cash market volume, at 53%, up from 14% in 2010. It is even more noticeable in more dynamic segments. About 70% of all Future and Options (F&O) trades and 73% of transactions in the stock futures segment are executed through automated trading. These figures indicate that human decisions are no longer behind the modern Indian market, but rather an environment where quick computer programs are in control.
The change from manual trading to computer trading has also altered the way in which market abuse occurs, making it the ‘mechanism of autonomous algorithmic manipulation’. Rather than secret human manipulation, computer programs now manipulate prices. The most popular techniques are spoofing and layering. Spoofing occurs when a trading program places numerous fake orders on one side of the order book to make it appear there is high demand for a stock. This is a created demand, which pushes prices up and down. When the price moves and other traders follow, the program automatically removes the fake orders and executes a real trade at the opposite end to make a fast profit. Layering is a similar approach, but with fake orders placed at several different price levels to manipulate the market depth. These computer programs make trades and process market data in milliseconds, so they can carry them out completely on their own by using these manipulative patterns. They do this only to maximise a pre-programmed mathematical goal without any learning or being aware of it throughout the transaction.
Legal Limits of the PFUTP Framework
The SEBI PFUTP Regulations, 2003, are the main instrument to confront market abuse in India. But this structure was designed to handle human wrongdoing, and there are three significant legal and practical restrictions in its use to deal with automated computer manipulation.
- The Human-Centric Nature of the Law: The first restriction is that the PFUTP is a human-centric approach to the law. Regulations 3 & 4 make this assumption: Market manipulation begins with someone who has a goal and is in communication with others to execute the goal. For the traditional human traders, where there is a clear link between human intention and a physical trade, this assumption truly holds. But automated systems don’t work this way. Instead, the programmer states the general objective, such as profit maximisation, and the software works out the best way to achieve the objective while adapting to the real-world information. If the software determines that the best method of profiting is to take orders and cancel them quickly, it will do that. Therefore, it is difficult to make any person liable under the law since this particular trade did not come from any human intent. The human-centric definition in the PFUTP is difficult to apply to the case where the decision is made solely by self-learning software.
- The Problem of Intent (Mens Rea) and Legal Attribution: The second restriction is the mens rea, or the problem of intent, as it is called in Indian securities law. The Supreme Court of India has expressed varying opinions with regard to the requirement of proof of intent in cases of market abuse over the years. In SEBI v. Shri Kanaiyalal Baldevbhai Patel (2017), the Court adopted a liberal interpretation, holding that ‘unfair trade practices’ under PFUTP do not require proof of bad faith or dishonesty if the action affects market transparency. This was further elaborated in SEBI v. Rakhi Trading Pvt. Ltd. (2018), where the Court concluded that the mere repetition of non-genuine transactions without any economic justification amounts to an unfair trade practice even without a direct intent to manipulate the market.
However, this objective, outcome-based approach exists alongside a far more stringent circumstantial standard laid down in SEBI v. Kishore R. Ajmera (2016) and Chintalapati Srinivasa Raju v. SEBI (2018). Dealing directly with fraudulent trade practices under PFUTP, the Court held that inferring market manipulation from circumstantial evidence requires a continuous, cogent and unbroken chain of events. This demonstrates a broader judicial scepticism toward fixing liability purely on suspicious trading patterns without concrete proof, a caution also reflected in Balram Garg v. SEBI (2022) in the insider trading context.
This doctrinal tension reveals that the core legal hurdle in AI-driven manipulation is not merely proving mens rea but a dual problem of intent and attribution. While Rakhi Trading permits SEBI to target manipulative outcomes, an autonomous algorithm breaks the causal chain required to attribute those outcomes to a legal entity. If SEBI proceeds under Kishore Ajmera, it can’t establish subjective intent because an algorithm lacks a legal mind and human developers did not explicitly order the manipulative trade. Conversely, if SEBI relies on Rakhi Trading, holding developers or brokers liable purely for autonomous algorithmic outputs risks imposing strict liability without a clear statutory framework for corporate attribution. Thus, the PFUTP framework struggles because it lacks a coherent mechanism to attribute autonomous algorithmic behaviour to an accountable human actor. - Surveillance and Due Process Challenges: The third restriction is that when SEBI’s technology detects suspicious activity, how do you get from detection to proof in court? In order to capture real-time trading through SEBI, they implemented the Integrated Market Surveillance System (IMSS). The IMSS is highly effective at identifying unusual behaviours such as quick order cancellations. A pattern, however, is not the same as a violation of the law.
When the market is volatile or when orders are cancelled or withdrawn because of the quick changes in the market, this is not manipulation; this is a legitimate and rational action taken by the automated system. Moreover, SEBI has not disclosed the functioning of its surveillance software, making accused traders vulnerable to transparency issues. They are unable to defend themselves easily because they do not understand the reason for their flagging, which denies them the right to be heard (audi alteram partem). Therefore, the surveillance technology is capable of identifying unusual patterns but not necessarily evidence of a legal violation, which raises fair trade concerns.
The Path to Regulatory Reform
To fix these legal loopholes, SEBI should revise its approach to effectively regulating automated trading. SEBI should take three focused decisions to reform the old rules to suit the new technology instead of trying to fit the old rules into the new technology.
First, rather than presenting a radical departure from existing doctrine, SEBI should formally codify the emerging ‘pattern plus effect’ standard into the PFUTP framework. Given the reasoning in Kanaiyalal Patel and Rakhi Trading, adopting this test should be understood as a clarification and statutory codification of an already emerging effects-based jurisprudence rather than a complete doctrinal shift. As the Supreme Court has already recognised that conduct distorting market integrity constitutes an unfair trade practice regardless of direct subjective intent, codifying this test brings much-needed regulatory certainty. Under this codified test, two objective conditions would be formally established: a pattern of trading (like repetitive non-bona fide orders) and an artificial effect on the market (like abnormal price changes or depth distortion). Establishing these would create a rebuttable presumption of market manipulation, shifting the burden to the market participant to prove a legitimate economic rationale through clear safe harbour provisions. Presenting this reform as a formalization of principles already embedded in Kanaiyalal and Rakhi Trading harmonises the proposal with existing jurisprudence, making it legally sound and readily adoptable by regulators and courts.
Second, the law must clearly set responsibility for the harm done by a computer program. This is a significant step in that direction, as SEBI’s algorithmic trading framework is set to be fully implemented from April 2026. The new rules make the broker the principal and the software provider the agent. That implies that brokers are accountable and strictly liable for any disruption of the market that results from the code that is running on their systems. The idea is similar to a European approach, MiFID II, which leaves investment firms accountable for their automated systems. The brokers are now held liable for the software’s results, which means that brokers will have to pay attention to their software and have emergency kill switches to immediately stop any malfunctioning software from running.
Third, the regulatory system needs to become more complicated with the increased complexity of the trading program being employed. This is addressed in the April 2026 framework by separating out two different types of automated systems. Simple systems involve some transparent, hard-coded logic that is easy to validate, such as automated average price execution models. Complex systems have self-learning logic, which can evolve over time and thus are more difficult to trace. According to SEBI’s recommendations, entities that offer such complicated products, which must be registered as a Research Analyst, and detailed reports must be maintained for each strategy. To implement a self-learning system, the system needs to be registered accordingly, and detailed reports need to be maintained for each strategy. Any change in the core logic of a complex system requires it to be registered as a new strategy. This will allow the regulators to keep a close watch on systems that are high-risk and self-learning, while not overburdening simpler, safer trading tools with compliance requirements.
Conclusion
Automated trading has transformed the way in which markets are manipulated, as it has freed up trading from humans and purpose. The SEBI’s PFUTP Regulations have been designed to deal with human wrongdoing and are not easily capable of dealing with autonomous software. Current legislation is based on intent to make a point; this approach is problematic when it comes to holding automated systems accountable. SEBI is required to update its regulations to ensure a fair and level playing field for investors. A shift from intent-based to a pattern-based standard, with transparent software classification and strict broker accountability, will help keep India’s capital markets safe and transparent in the digital era.
Sumit is a third Year Law Student at National Law University, Odisha.
