Artificial Intelligence is now everywhere in the Indian securities market. Open any stock market app today and you will find AI-powered screeners, AI-generated trading alerts, robo-advisory tools, automated portfolio suggestions, and even chatbots answering investment queries in real time. Many fintech platforms proudly market themselves as “AI-driven.”
But this rapid adoption raises a very important legal and compliance question:
If AI gives investment advice and something goes wrong, who is responsible?
Can SEBI regulate advice generated by a machine? And more importantly, can an AI system itself ever be regulated, licensed, or held accountable under Indian securities law?
These are no longer futuristic questions. SEBI has already started building a regulatory framework around AI usage in the securities market, and firms using AI in advisory or research functions need to pay attention now.
SEBI Has Already Started Acting on AI
In recent years, SEBI’s approach toward AI has moved from observation to active regulation.
In January 2025, SEBI introduced an important requirement for Research Analysts (RAs): firms must now disclose to clients whether AI tools are being used in their research process and to what extent.
Shortly after that, SEBI also clarified that Investment Advisers cannot escape liability by blaming algorithms or AI systems. If AI-generated advice turns out to be misleading or inaccurate, the registered adviser remains fully responsible.
Then came SEBI’s major milestone — the June 2025 Consultation Paper on “Responsible Usage of AI/ML in Indian Securities Markets.”
This consultation paper made SEBI’s position very clear:
Using AI does not reduce regulatory responsibility. It increases it.
The regulator proposed a governance framework based on principles like accountability, transparency, fairness, auditability, ethics, and data privacy.
In simple terms, SEBI wants firms to ensure that AI systems are not operating like black boxes without human oversight.
The Real Legal Problem: AI Cannot Be Registered With SEBI
The biggest challenge is actually very basic.
Under existing laws such as:
- SEBI (Investment Advisers) Regulations, 2013
- SEBI (Research Analysts) Regulations, 2014
investment advice must come from a registered human professional.
A Research Analyst or Investment Adviser must satisfy qualification criteria, obtain NISM certifications, maintain compliance records, and remain answerable to SEBI.
But an AI model cannot:
- obtain NISM certification,
- register with SEBI,
- sign a research report,
- appear before SEBI,
- or face suspension or penalties.
And that creates what many experts now call the “accountability gap.”
AI can generate recommendations, but legally, only a human can be held responsible.
Because of this, SEBI’s current approach is straightforward:
Any AI-generated advice remains the responsibility of the registered human professional associated with it.
Whether the advice comes from a human analyst or from an AI engine running behind the platform, the accountability remains with the regulated entity and its authorised professionals.
Why This Matters for Fintechs and Advisory Platforms
This issue is especially important for:
- Registered Investment Advisers (RIAs)
- Research Analysts (RAs)
- Fintech platforms
- Algo trading businesses
- Robo-advisory firms
- Wealth-tech startups
- AI-powered stock advisory apps
Many firms today use third-party AI tools or APIs without fully understanding the compliance implications.
For example:
- What happens if an AI chatbot recommends buying a stock?
- What if an AI screener generates a personalised portfolio?
- What if a machine learning model gives inaccurate predictions leading to investor losses?
SEBI is increasingly likely to treat these outputs as regulated investment advice — even if the platform describes them merely as “educational” or “informational.”
The regulator’s enforcement trend already suggests this broader interpretation.
For more details, go to www.kpregtech.com
Three Major Compliance Risks Firms Must Understand
1. Liability for Third-Party AI Tools
A large number of fintech companies do not build their own AI systems. They rely on external vendors or integrate AI APIs into their platforms.
But under SEBI’s emerging framework, using a third-party AI tool does not shift responsibility.
If the AI generates misleading advice, the regulated entity may still remain liable.
This means firms must carefully review vendor agreements and ensure proper clauses exist around:
- data quality,
- model accuracy,
- audit rights,
- confidentiality,
- and allocation of liability.
Simply relying on standard software terms and conditions may not be enough.
2. What Exactly Counts as “Investment Advice”?
– This is one of the biggest grey areas.
Suppose an AI chatbot answers:
“This stock looks undervalued for long-term investment.”
Q.1 Is that investment advice?
Ans. What if an AI engine suggests:
“Based on your risk profile, allocate 40% to equities and 60% to debt.”
Q.2 Does that require RIA registration?
Ans. SEBI has not yet issued a definitive technical test for AI-generated interactions. However, recent enforcement actions indicate that the regulator is likely to adopt a very broad interpretation.
If an AI-generated output can reasonably influence an investor’s decision to buy, sell, or hold securities, SEBI may treat it as regulated advice.
3. The “Black Box” Problem
One of SEBI’s strongest expectations is that firms should be able to explain how their AI systems arrive at recommendations.
This is called traceability or auditability.
For simple rule-based systems, this is manageable.
But modern AI systems — especially large language models and advanced machine learning tools — often function as black boxes. Even developers may not fully understand how certain outputs are generated.
This creates a difficult compliance challenge.
If a regulator asks:
“Why did your AI recommend this stock to this client?”
- the firm must be able to provide a reasonable explanation.
As SEBI’s framework evolves, firms using opaque AI systems without adequate documentation or oversight may face higher regulatory risk.
What Firms Should Start Doing Immediately
Even though SEBI’s final AI framework is still evolving, firms should not wait for mandatory enforcement before preparing.
Some practical steps include:
Clearly disclose AI usage
Clients should know:
- whether AI is being used,
- where it is being used,
- and what human oversight exists.
Generic disclosures about “technology-enabled services” are unlikely to satisfy regulatory expectations.
Ensure human accountability
Every AI-generated recommendation should ultimately be reviewed, owned, or supervised by a named registered professional.
There should never be uncertainty about who is responsible for a client-facing output.
Create an internal AI governance policy
Firms should adopt a documented AI governance framework covering:
- model testing,
- approval processes,
- monitoring,
- escalation mechanisms,
- human review procedures,
- and error management.
Review all AI touchpoints
Many firms underestimate how deeply AI is embedded in their operations.
A proper compliance review should map every AI interaction, including:
- chatbots,
- portfolio recommendation engines,
- automated alerts,
- screening tools,
- risk profiling systems,
- and personalised notifications.
SEBI Is Also Using AI for Enforcement
An interesting development in 2026 is that SEBI itself has reportedly started using AI tools to monitor:
- financial influencers,
- social media investment content,
- advertisements,
- and potentially misleading financial promotions.
This means detection capabilities are becoming stronger.
Non-compliant AI-generated advice is now more likely to be identified quickly — especially on digital platforms.
Global Regulators Are Moving in the Same Direction
India is not alone in facing this issue.
Regulators across the world are taking similar positions.
- In the United States, FINRA and the SEC have clarified that firms remain responsible for AI-generated recommendations.
- Singapore’s MAS has stressed explainability, fairness, and human accountability.
- Australia’s ASIC has also focused on governance and oversight obligations.
SEBI’s framework aligns closely with these international trends.
However, India’s model appears stricter in one key aspect:
Responsibility is tied very specifically to the registered human professional.
That makes accountability far more personal.
Conclusion
SEBI may not regulate AI as an independent entity yet — but it is certainly regulating the use of AI in investment advisory and research activities.
The core principle is already clear:
AI cannot replace regulatory accountability.
If AI-generated advice reaches a client, a registered human professional and the regulated entity remain responsible for it.
For fintech companies, RIAs, Research Analysts, and AI-driven advisory businesses, this is no longer a future issue. It is an active compliance priority.
The firms that build strong AI governance structures today will be far better prepared for the regulatory environment that is clearly approaching.
Author Bio
