Loading live prices…
Regulation

India Sets Audit Rules for AI Agents as AI-Assisted Hackers Hit Crypto Exchanges

India has moved to put formal checks around one of the fastest-growing risks in modern finance: artificial intelligence systems that can plan, decide and act without a human in the loop. In the week beginning November 8, 2026, the CERT-In audit rules for autonomous AI agents were published, according to a report by TokensTimes. The timing was pointed. The same week, CERT-In also named AI-assisted hackers hitting fintech companies and crypto exchanges, drawing a direct line between the technology that is transforming finance and the technology that is now being used to attack it. The India AI agent audit rules arrive at a moment when AI agents are no longer a laboratory curiosity. They are being given spending power, access to accounts, and the ability to sign transactions, and regulators in India have decided that such power needs to be checked, logged and tested before it causes harm.

Advertisement

It is important to be clear about what is known and what is not. The TokensTimes report that forms the basis of this article is short on detail. It confirms that CERT-In has published audit rules for AI systems that plan and act on their own, and that the agency has separately drawn attention to AI-assisted hackers targeting fintechs and crypto exchanges in the same week. It does not, in the material supplied for this article, spell out the full text of the rules, the timetable for compliance, the penalties for failure, or the precise technical standards that auditors will apply. That thinness is itself part of the story. Regulators often announce a framework first and fill in the technical detail later, through guidance, consultations and enforcement. What follows in this article therefore separates the confirmed facts from the wider background: the facts are the publication of the rules and the warning about AI-assisted attacks, while the analysis of what audits usually involve, why autonomous agents worry regulators, and how India fits into a global pattern is context that helps readers understand why the announcement matters, even before the fine print is public.

What CERT-In’s audit rules mean for AI agents

CERT-In, formally the Indian Computer Emergency Response Team, is the agency that coordinates how India responds to cyber incidents, issues advisories, and sets expectations for how organisations protect their systems. When it publishes audit rules for AI agents, it is saying that organisations using autonomous systems in sensitive areas should be able to prove, to an independent checker, that those systems behave safely and predictably. An audit in this sense is not a single test. It is a structured review of how an agent was built, what data it was trained on, what limits were placed on it, how its decisions are logged, and how it responds when something goes wrong. For a crypto exchange, that could mean showing that a trading agent cannot move funds beyond a set limit, cannot approve a withdrawal without a second check, and cannot be tricked by a fake instruction hidden in a message, a document or a website.

Why regulators are moving on autonomous agents now

Regulators are moving now because AI agents have moved from demonstration to deployment in a very short time. Only a few years ago, the idea of an AI system that could manage a portfolio, negotiate a contract, or run a customer service operation on its own sounded futuristic. Today, banks, fintechs and crypto firms are testing agents for research, trading, compliance checks, customer support and back-office work. The technology has also become cheaper and easier to use, which means smaller firms, not just the largest banks, can deploy it. That spread is good for innovation, but it also means that a failure in one small firm can ripple through the wider system if that firm holds customer funds or connects to a larger exchange.

Advertisement

AI-assisted crypto hacks and the new attacker

The other half of the week’s news is the warning about AI-assisted hackers hitting fintechs and crypto exchanges. AI-assisted crypto hacks are not a single technique. They describe a range of ways in which attackers use artificial intelligence to make their attacks faster, more convincing and harder to detect. An attacker might use AI to write phishing emails that are perfectly tailored to a specific employee, to generate deepfake voice or video messages that impersonate a senior executive, to scan code for vulnerabilities far more quickly than a human could, or to automate the process of trying stolen credentials across dozens of exchanges. The result is that attacks that once required a skilled team and weeks of preparation can now be launched more cheaply and at greater scale.

What an AI agent audit looks like in practice

Although the full detail of the India AI agent audit rules has not been published in the material available for this article, audits of this kind generally follow a familiar pattern, and it is useful to understand what firms are likely to be asked to do. The first step is usually an inventory: what agents are in use, what can each one do, what systems can it access, and what limits apply to it. Many organisations do not have a complete list, because agents have been adopted team by team, without central oversight. An audit forces that list to be created, which is often the first time senior managers realise how many autonomous systems are already running inside their firm.

India in a global pattern of crypto and AI regulation

In the United States, the Securities and Exchange Commission has been working on a framework for crypto assets, often described as Regulation Crypto Assets, which aims to clarify when a token is a security and what disclosures and protections should apply. The US debate has been more contested, with arguments about whether existing securities law already covers crypto, and about how to avoid stifling innovation while protecting investors. The US experience also shows why the Indian approach to AI agents matters: when regulation is unclear, firms either hold back from building useful products or build them in ways that are hard to supervise. Clear audit rules, even if they are demanding, give firms a target to aim for.

What this means for exchanges, fintechs and users

For crypto exchanges operating in India or serving Indian users, the immediate implication is that AI agents can no longer be treated as an internal experiment. If an exchange uses an agent to handle customer queries, monitor transactions, or execute trades, it should expect to be asked how that agent is tested, what it can access, and how its actions are logged. Exchanges that have already built those controls will find the audit process straightforward. Those that have deployed agents quickly, without a clear inventory or a set of limits, may need to slow down and build the governance around the technology before they expand its use. That is a cost, but it is also a form of protection: a firm that can show it audited its agents will be in a stronger position if something goes wrong.

Advertisement

For fintechs, the lesson is similar, with an added emphasis on the human side. Many AI-assisted attacks succeed not because the technology is broken, but because a person is tricked. A deepfake call from a supposed executive, a perfectly written email from a supposed regulator, or a message that appears to come from a trusted colleague can persuade an employee to approve a payment or share a credential. Audits that test only the software will miss that risk. Firms therefore need to train their staff to be sceptical of urgent requests, to verify instructions through a second channel, and to report suspicious messages quickly. The audit rules, if they follow the usual pattern, will ask about that training as well as about the code.

Frequently asked questions

What are the India AI agent audit rules?

The India AI agent audit rules are audit requirements published by CERT-In for AI systems that plan and act on their own, known as autonomous AI agents. They set expectations that organisations using such agents in sensitive areas, including finance and crypto, should be able to show that the agents are tested, limited, logged and overseen by a responsible person.

What is CERT-In?

Advertisement

CERT-In is India’s national cyber agency, formally the Indian Computer Emergency Response Team. It coordinates responses to cyber incidents, issues security advisories, and sets expectations for how organisations protect their systems and report serious incidents.

What are AI-assisted crypto hacks?

AI-assisted crypto hacks are attacks in which hackers use artificial intelligence to make their methods faster, more convincing or harder to detect. That can include AI-written phishing messages, deepfake impersonation, automated scanning for software flaws, and automated attempts to use stolen credentials across many exchanges and fintechs.

Do the audit rules ban AI agents in crypto?

Advertisement

No, based on the information available, the rules do not ban AI agents. They require firms to audit and control them. The approach is to set guardrails and require testing, so that innovation can continue inside a framework that protects users and the wider financial system.

How do India’s rules compare with the UK and US?

India’s move fits a global pattern. The UK’s FCA has consulted on the crypto perimeter and DeFi, and the US SEC has worked on a framework for crypto assets. All three approaches share a focus on evidence, testing and clear responsibility, even though the legal details and timetables differ.

Source: TokensTimes

Related reading on MM Crypto Analysis Web

More in Regulation on MM Crypto Analysis Web.

Advertisement
Advertisement

MM Crypto Team

MM Crypto Team covers crypto news, prices and market analysis.

One thought on “India Sets Audit Rules for AI Agents as AI-Assisted Hackers Hit Crypto Exchanges”

Leave a Reply

Your email address will not be published. Required fields are marked *

Advertisement
Advertisement
Advertisement