Understanding AI and Indian Consumer Protection Laws
Covering
- How AI is currently harming consumers in India, including misleading product images, unreliable mass‑produced content, deceptive online profiles, and automated health guidance without adequate safeguards.
- Who bears responsibility when AI causes harm, the paper frames the central accountability question across the user of the tool, the AI provider, and the publishing platform.
- A clear framework of four consumer harm patterns, identity/appearance deception, quality/content deception, manipulation of consumer decisions, and automated decisions without meaningful human oversight.
- How Indian consumer law applies to AI disputes, unfair trade practices, deficiency of service, product liability principles, and the enforcement role (and current under‑use) of the Central Consumer Protection Authority (CCPA).
- What online platforms must do under digital and e‑commerce rules, “reasonable efforts” obligations under the E‑Commerce Rules, safe‑harbour and due‑diligence under the IT Act/Intermediary Rules, and the 2026 SGI framework for disclosure, verification, labelling, and rapid removal on large platforms.
- Ten case studies, legal gaps, and practical reforms, real‑world scenarios across sectors,
Case Study | Topic Covered |
|---|---|
Dating App | AI-generated or altered profile photos creating deceptive identity representation on dating platforms. |
Online Marketplace | AI-generated product images misrepresenting the quality or appearance of goods sold online. |
News Website | AI-written articles presented as human-authored content despite factual errors and non-disclosure. |
Job Portal | Platform-generated AI CVs inflating candidate profiles and distorting recruitment decisions. |
Real Estate Platform | AI-generated property images misrepresenting the condition, size, or existence of listed properties. |
Health App | AI-generated medical guidance creating unsafe reliance without adequate validation or human oversight. |
Matrimonial Platform | AI-generated photos and profiles enabling fraud on services marketed as verified match-making platforms. |
Food Delivery Platform | AI-generated food images misleading consumers about quality, ingredients, and actual presentation. |
Investment Platform | AI-generated financial advice creating loss through outdated, unsuitable, or misleading recommendations. |
Online Education Platform | AI-generated course content and certifications undermining learning quality, competence, and trust. |
1. What Is Artificial Intelligence And Why Should Consumers Care?
Artificial Intelligence, or AI, refers to computer systems that can perform tasks that normally require human thinking. This includes writing text, generating photographs, recognizing faces, making decisions, and having conversations. AI is not new, but in the last few years it has become extremely powerful and extremely cheap to use.
Today, anyone with a smartphone and a free account on an AI tool can generate a completely fake but realistic-looking photograph in seconds. Anyone can produce a 2000-word article that reads like it was written by an expert. Anyone can create an AI voice that sounds exactly like a real person.
These capabilities are exciting for many uses. But they also create serious problems for consumers.
When an AI-generated photo, article, or recommendation causes you harm, who is responsible? Is it the person who used the AI tool? Is it the company that made the AI? Or is it the platform that published the content and allowed it to reach you?
The Key Question
The answer to that question is what consumer protection law is supposed to provide. Unfortunately,
India's current laws were written before powerful AI tools existed. They do not give a clear or complete answer. This paper examines what the law currently says, where it falls short, and what
needs to happen to properly protect Indian consumers in the age of AI.
2. The Scale Of The Problem In India
India has over 900 million internet users. It is one of the fastest-growing digital economies in the world. E-commerce platforms process hundreds of millions of transactions every year. Dating apps have tens of millions of active users. News and content platforms publish millions of articles daily.
Financial platforms give investment advice to millions of people.
Across all of these, AI is now generating content that consumers see and act upon. A buyer chooses a product based on a photo. A user goes on a date based on a profile picture. A reader forms an
opinion based on an article. An investor makes a decision based on a recommendation. If any of that content is AI-generated and misleading, the consumer suffers real harm.
The problem is not hypothetical. Studies have found that nearly one in five product reviews on major
Indian shopping platforms show signs of AI generation. Dating platforms are flooded with profiles using AI-altered or entirely AI-generated photographs. Content farms use AI to flood news websites with articles that appear credible but are factually wrong. All of this harms consumers, and all of it is happening right now.
3. The Four Types Of Harm That AI Creates For Consumers
It helps to organize AI-related consumer harm into four clear categories. Most of the case studies in this paper will fall into one or more of these categories.
Harm Type 1: Deception About Identity Or Appearance
This happens when AI is used to create or alter images, videos, or profiles to misrepresent who or what something is. A seller uses an AI-generated photo to make a cheap product look premium. A
dating app user uses AI to replace their real photo with an idealised one. A business uses an AI-generated video testimonial to fake customer endorsements.
Harm Type 2: Deception About Quality Or Content
This happens when AI generates text or content that claims to have expert quality but is actually unreliable. An AI-written medical article gives dangerously wrong advice. An AI-generated legal template is legally incorrect. A real estate listing uses AI-generated descriptions that exaggerate property features.
Harm Type 3: Manipulation Of Consumer Decisions
This happens when AI systems analyse consumer data and use it to manipulate choices. A shopping platform uses AI to show a consumer a higher price because the AI has calculated that this particular person is likely to pay more. A food delivery platform shows certain restaurants first because the AI is optimising for profit rather than consumer preference.
Harm Type 4: Automated Harm Without Human Oversight
This happens when AI makes consequential decisions about consumers without any human checking the outcome. A loan application is rejected by an AI without any human review. A social media account is banned by an automated system. An insurance claim is denied by an AI model that contains a bias built into its training data.
What The Law Says
The Consumer Protection Act, 2019: Main Legal Provision
The Consumer Protection Act, 2019, is the most important law protecting Indian consumers. It replaced an older 1986 law and was designed to be more modern and more powerful. Let us look at its most important provisions and what they mean in simple terms.
What Is An Unfair Trade Practice?
Section 2(47) of the Act defines an unfair trade practice as any method a business uses to promote its goods or services that is unfair or deceptive. This includes making false claims about a product, using
misleading descriptions, and creating fake endorsements.
If a seller puts up a fake AI-generated photo of a product that makes it look much better than it is, and you buy it because of that photo, the seller has committed an unfair trade practice. This is illegal.
What Is Deficiency Of Service?
Section 2(11) defines deficiency of service as any fault, shortcoming, or inadequacy in the quality of a service that a business promised to provide. If a platform allows deceptive content that harms
consumers without putting any safeguards in place, that platform itself may be guilty of deficiency of service.
If a dating platform allows users to upload AI-generated profile photos without any detection system, and you pay for a premium subscription to meet real people, the platform has failed to deliver the service it promised. That is a deficiency of service.
Product Liability: Chapter Vi Of The Act
Chapter VI of the Consumer Protection Act, 2019 introduced no-fault product liability. This means that if a product is defective and causes you harm, the manufacturer or seller is responsible even if
they did not intend the product to be defective. You do not need to prove they were careless. You only need to prove the product was defective and that it caused your harm.
The big legal question with AI is: is AI a product or a service? If it is a product, no-fault liability applies. If it is a service, then the consumer must prove negligence or a contractual shortcoming. India has not yet answered this question clearly in law. This is one of the biggest legal gaps we discuss later in this paper.
The Central Consumer Protection Authority
The CCPA is a powerful regulatory body created by the 2019 Act. It has the power to investigate consumer complaints on its own without waiting for someone to file a case, issue directions to stop
unfair practices, recall defective products, and impose penalties. For a first violation, the penalty can be up to Rs. 10 lakhs. For repeat violations, it can reach Rs. 50 lakhs.
The CCPA has not yet actively exercised its powers against AI-driven consumer harm. This is a significant gap in enforcement. The authority exists. The will to use it has been slow to develop.
5. The E-Commerce Rules, 2020: Rules For Online Platforms
The Consumer Protection (E-Commerce) Rules, 2020 apply specifically to online marketplaces. They require e-commerce platforms to give consumers accurate information about products, prevent sellers from posting misleading content, take reasonable steps to ensure sellers comply with the law, and provide a proper grievance mechanism.
Rule 5 of these rules says every e-commerce entity must make reasonable efforts to ensure that product information on its platform is not false or misleading. If a platform allows sellers to upload
AI-generated product images without any verification, it is arguably violating this rule. The phrase
'reasonable efforts' is key. In 2026, deploying basic AI-detection technology is a reasonable effort for any large platform to make. Not doing so is a failure.
6. The IT Act, 2000 And The Intermediary Rules: Rules For Platforms
The Information Technology Act, 2000 was India's first major digital law. Section 79 of this Act gives online platforms something called safe harbour protection. This means platforms are generally not responsible for content posted by their users, as long as the platform meets certain conditions.
What Is Safe Harbour And When Do Platforms Lose It?
Safe harbour is like a legal shield. A platform cannot be held liable for every piece of content every user uploads, because that would make it impossible to run any online platform. But this shield comes with conditions. The platform must exercise due diligence in operating its services. If the platform has knowledge of illegal or harmful content and fails to act, it loses the shield and becomes responsible.
If a platform knows that sellers are posting fake AI-generated product images and does nothing about it, it loses its legal protection and can be held liable for the harm caused to consumers.
The 2021 Intermediary Guidelines
The IT (Intermediary Guidelines and Digital Media Ethics Code) Rules, 2021 set out the specific due diligence steps platforms must follow to keep their safe harbour protection. Platforms must publish clear policies about what content is allowed, take down harmful content when they receive complaints, resolve grievances within 15 days, and designate officers responsible for compliance.
The 2026 Amendment: The Most Important Update For AI
In February 2026, India made a major update to these rules. This is the most significant step India has taken so far to regulate AI-generated content. The amendment introduced the concept of
Synthetically Generated Information, or SGI. SGI means any audio, image, or video that was created or altered by AI in a way that makes it look or sound real.
Under these new rules, large platforms (those with more than 50 lakh registered users) must require users to declare if their content is AI-generated, verify those declarations using automated tools, label
AI-generated content clearly before it is published, remove harmful AI-generated content within 3 hours of receiving an order.
If a platform fails to do these things, it loses its safe harbour protection. This means it can be sued directly for the harm caused by AI-generated content on its platform.
8. India's AI Governance Guidelines, November 2025
In November 2025, the Ministry of Electronics and Information Technology released India's AI
Governance Guidelines. These guidelines outline India's approach to AI regulation. They are built around seven principles: Safety, Ethics, Transparency, Accountability, Fairness, Inclusion, and
Privacy.
These guidelines recommend an approach that is called light-touch regulation. This means the government wants to encourage AI innovation and not burden it with heavy rules. The guidelines rely mostly on existing laws and voluntary measures rather than creating new binding obligations.
India's AI Governance Guidelines are not law. They cannot be enforced by any court or regulator. If a company ignores them, there is no penalty. This is perhaps the single biggest gap in India's current AI regulatory framework.
Ten Case Studies
The following case studies are based on real patterns of conduct observed across various types of digital platforms in India. The case studies cover a wide range of platform types to show that AI-
driven consumer harm is not limited to any one industry.
CASE STUDY 01 The Dating App With AI-Generated Profile Photos | |
|---|---|
Platform Type | Social Networking and Dating Application |
The Situation | A popular paid dating application allows users to upload any photograph as their profile picture without running any verification or AI-detection check. Many users upload AI-generated photographs that make them appear significantly more attractive than they are in real life. Some users upload photographs of entirely different people. Subscribers who pay a monthly fee to find genuine connections meet people who look nothing like their profiles. Some users pay for premium meet-up features and travel to other cities for dates, only to find that the person they were matched with bears no resemblance to the profile. |
Consumer Harm | Financial loss (subscription fees, travel costs), emotional distress, waste of time, and in some cases safety risks from meeting unknown individuals under false pretences. The consumer paid for a service that promised authentic connections and received a fundamentally false experience. |
Applicable Laws | Section 2(47) CPA 2019: Unfair Trade Practice (false representation of service quality). Section 2(11) CPA 2019: Deficiency of Service (platform failed to provide what was promised). IT Rules 2026: Failure to detect and label AI-generated profile images before publication. DPDP Act 2023: Potential misuse of user data processed to generate match recommendations. |
Legal analysis | The platform is liable for deficiency of service. It took subscription fees promising a genuine matching experience and delivered a fake one. The failure to deploy AI-image detection for profile photos, which is technically feasible and commercially available, constitutes a clear failure of due diligence under Section 79 of the IT Act and the 2026 Intermediary Rules Amendment. The consumer is entitled to a refund of subscription fees and compensation for provable losses. |
Case Study 02 The Online Marketplace With AI-Generated Product Images
Platform Type | E-Commerce Marketplace |
|---|---|
The Situation | A large e-commerce platform allows third-party sellers to list products and upload their own product images. Sellers use AI image-generation tools to create photorealistic images of goods they do not actually have or that look far superior to the real product. A consumer orders a piece of furniture that appears to be solid wood from a beautifully lit AI-generated photograph. The delivered product is made of low-quality particle board. The platform processed the listing without running any check on whether the image was AI-generated or whether it accurately represented the product. |
Consumer Harm | The consumer spent money on a product that was misrepresented. The product received is materially different from what was depicted. The consumer must now spend money and time to return the product, file a complaint, and wait for a refund that may or may not arrive. The consumer trusted the platform and was deceived. |
Applicable Laws | Section 2(47)(i) CPA 2019: Unfair Trade Practice by seller through false representation of product quality. E-Commerce Rules 2020 Rule 5: Platform's failure to take reasonable steps to ensure product information is not misleading. Section 2(11) CPA 2019: Deficiency of Service by the platform. IT Rules 2026: Failure to implement pre-publication checks for AI-generated imagery. |
Legal analysis | Both the seller and the platform are liable. The seller is primarily liable for the deceptive listing. The platform is jointly liable for failing to detect and remove AI-generated deceptive content, especially given that this is a pattern of conduct on the platform and not an isolated incident. The platform cannot hide behind its terms of service to avoid liability when it has systemically failed its duty of care toward consumers. |
CASE STUDY 03 The News Website With AI-Written Articles | |
|---|---|
Platform Type | Online News and Content Publishing Platform |
The Situation | A news and information website publishes hundreds of articles daily. To reduce costs, the website uses AI to generate articles on health, finance, and legal topics. The AI-generated articles appear under human author names. They often contain factual errors, outdated information, or misleading advice. A reader relies on an AI-generated article about drug interactions that contains incorrect medical information. Acting on that advice causes the reader a health problem. The website never disclosed that the article was AI-generated. |
Consumer Harm | Physical harm from acting on incorrect AI-generated health advice. Financial loss if the same pattern applies to financial advice articles. Psychological harm and loss of trust. The reader had a reasonable expectation that articles published under a human author's name were researched and verified by a human being. |
|---|---|
Applicable Laws | Section 2(47) CPA 2019: Unfair Trade Practice through false representation (articles presented as human-written when AI-generated). Section 2(11) CPA 2019: Deficiency of Service (the service of providing reliable information was not delivered). Press Council guidelines on journalistic standards. IT Rules 2026: SGI disclosure obligations apply to AI-written text content on major platforms. Potential negligence under tort law for physical harm caused by incorrect health information. |
Legal analysis | The publishing platform is liable for both the deceptive non-disclosure and the consequential harm. Publishing AI-generated content under human author names is a form of deception. If the platform's website has more than 50 lakh users, the 2026 IT Rules require disclosure of AI-generated content. The absence of such disclosure, combined with harm to a reader, creates a strong case for compensation under the Consumer Protection Act and potentially under ordinary negligence principles. |
CASE STUDY 04 The Job Portal With AI-Generated Candidate Profiles | |
|---|---|
Platform Type | Online Employment and Recruitment Platform |
The Situation | A job recruitment platform allows job seekers to create profiles that are enhanced by an AI tool built into the platform itself. The AI tool rewrites resumes to make them appear more impressive, adds skills the candidate has not listed but statistically matches to the role, and in some cases generates professional profile photographs. Employers paying the platform for access to verified candidate talent pools are making hiring decisions based on AI-augmented profiles. They hire candidates who do not have the skills attributed to them by the AI, causing business disruption and financial loss. |
Consumer Harm | Employers who pay for premium recruitment services receive inaccurate candidate information and make bad hiring decisions. Job seekers who do not use the AI enhancement tool are at a competitive disadvantage due to the platform's own AI inflation of competitor profiles. The trust between employer, candidate, and platform is systemically undermined. |
Applicable Laws | Section 2(47) CPA 2019: Unfair Trade Practice by the platform through its AI tool that generates false candidate credentials. E-Commerce Rules 2020: Platform's failure to ensure accuracy of information on its service. |
Section 2(11) CPA 2019: Deficiency of Service to employers who paid for verified talent access. DPDP Act 2023: Processing of candidate personal data to generate fabricated professional attributes without explicit consent. | |
Legal analysis | The platform is directly liable, not merely as a facilitator but as the creator of the deceptive content through its own AI tool. This removes any ambiguity about who is responsible. The employer-consumers who subscribed to the platform's services have a clear case for refund and compensation. The platform's AI enhancement feature, if it generates information the candidate did not provide, is a defective service within the meaning of the Consumer Protection Act. |
CASE STUDY 05 The Real Estate Platform With AI-Generated Property Photos | |
|---|---|
Platform Type | Online Property Listing and Real Estate Platform |
The Situation | A property listing website allows landlords and real estate agents to upload photographs of properties available for rent or sale. Many listers use AI image-generation tools to create photographs showing clean, bright, spacious interiors. The actual properties are in poor condition, much smaller, or have structural problems not visible in the AI images. Home-seekers pay booking fees or advances based on these photographs, sometimes without seeing the property in person, and arrive to find a completely different reality. In some cases, the property depicted in the AI images does not exist in the location advertised. |
Consumer Harm | Financial loss (advance payments, brokerage fees, moving costs), emotional distress, homelessness risk when a consumer relocates expecting a property that does not match the listing, and potential safety issues if structural defects were concealed by AI imagery. |
Applicable Laws | Section 2(47)(i) CPA 2019: Unfair Trade Practice through false representation of property condition and appearance. Section 2(11) CPA 2019: Deficiency of Service by the platform. Real Estate (Regulation and Development) Act, 2016: Obligations on real estate agents to provide accurate information. E-Commerce Rules 2020: Platform obligation to prevent misleading product (property) information. IT Rules 2026: Obligation to detect and label AI-generated images on large platforms. |
Legal analysis | Both the individual lister and the platform carry liability. The lister is guilty of an unfair trade practice. The platform, by failing to implement image verification, has been deficient in its service and has violated its obligations under the E-Commerce Rules. Consumers who paid advances are entitled to full refunds. Consumers who suffered additional losses due to relocation expenses have grounds for consequential compensation. |
CASE STUDY 06 The Health App With AI-Generated Medical Advice | |
|---|---|
Platform Type | Digital Health and Wellness Application |
The Situation | A health and wellness app offers personalised health advice, diet plans, and symptom analysis. The app uses an AI model that generates responses to user health queries. The AI has not been validated by any medical authority. It generates confident-sounding advice that is sometimes dangerously incorrect. A user experiencing chest pain inputs symptoms into the app. The AI recommends rest and antacids. The user delays seeking emergency medical care. The delay results in a serious cardiac event that could have been better managed with timely intervention. The app's terms and conditions state in small print that it does not provide medical advice, though its interface is designed to resemble a doctor consultation. |
Consumer Harm | Serious physical harm or death resulting from delayed medical intervention caused by incorrect AI-generated health advice. The consumer subscribed to a service that was designed and marketed to feel like a reliable health advisor. |
Applicable Laws | Section 2(47) CPA 2019: Unfair Trade Practice through deceptive interface design that implies medical reliability. Section 2(11) CPA 2019: Deficiency of Service for providing unreliable health guidance. Indian Medical Council Act: Restrictions on entities providing medical advice without proper authorisation. Consumer Protection Act Section 86: Product seller liability if the AI tool is treated as a product. Tort law: Negligence for failure to include adequate safety warnings and human oversight mechanisms. |
Legal analysis | The disclaimer in the terms of service does not eliminate liability when the service is designed to mimic medical consultation. Courts in India have consistently held that businesses cannot use fine-print disclaimers to escape liability for fundamental deceptions in their service design. The app developer is liable for deficiency of service and potentially for negligence if the AI advice was the proximate cause of harm. This is an area where product liability principles, applied to AI as a product, would provide the strongest remedy. |
CASE STUDY 07 The Matrimonial Platform With Fake Profiles | |
|---|---|
Platform Type | Online Matrimonial and Match-Making Service |
The Situation | A matrimonial website charges families and individuals substantial annual fees to access what it describes as a verified database of genuine marriage prospects. The platform allows profile photographs to be uploaded without any AI-detection system. Many profiles use AI-generated photographs of idealised appearances. Some profiles are entirely fabricated, created by scammers using AI-generated photos, AI-written bios, and false income and education credentials. Families spend months communicating with fake prospects, invest emotional energy in what they believe are genuine marriage negotiations, and in some cases transfer money based on fabricated circumstances. |
|---|---|
Consumer Harm | Financial fraud in cases involving money transfer to scammers. Severe emotional distress for families investing in marriage negotiations with non-existent individuals. Privacy violations when personal family information is shared with fraudulent profiles. The stakes in matrimonial fraud are uniquely high given the social and emotional significance of marriage in Indian society. |
Applicable Laws | Section 2(47) CPA 2019: Unfair Trade Practice (false representation of a verified database). Section 2(11) CPA 2019: Deficiency of Service (verification promised was not performed). IT Rules 2026: Obligation to verify and label AI-generated profile content on large platforms with over 50 lakh users. Bharatiya Nyaya Sanhita 2023: Provisions relating to fraud and cheating applicable to scammers. DPDP Act 2023: Obligation to protect user personal data shared with verified profiles. |
Legal analysis | The platform bears heavy liability because it specifically marketed its service as offering verified profiles. This is not a case of passive hosting. The platform actively made a promise of verification. The failure to implement basic AI-image detection and profile verification constitutes a direct breach of the advertised service. Families who paid subscriptions and suffered harm are entitled to refunds, compensation, and access to platform data to support any fraud complaint against individual scammers. |
CASE STUDY 08 The Food Delivery Platform With AI-Generated Food Photos | |
|---|---|
Platform Type | Online Food Ordering and Delivery Service |
The Situation | A food delivery platform allows restaurants to upload photographs of their dishes. Many restaurants, particularly smaller establishments, use AI image-generation tools to create photographs of food items that look nothing like the actual dish. A customer orders a biryani based on a photograph showing a richly coloured dish with large pieces of meat, fresh garnish, and premium ingredients. The delivered dish is a pale, sparsely |
garnished product that bears no resemblance to the photograph. This is a systematic pattern: AI-generated food images are being used across hundreds of restaurant listings on the platform to drive orders. | |
Consumer Harm | Financial loss from paying for food that is not as described. The experience of receiving a product that is materially different from what was ordered. Loss of trust in the platform's listings. For consumers with dietary restrictions or allergies, a mismatch between depicted and actual ingredients can also create health risks. |
Applicable Laws | Section 2(47)(i) CPA 2019: Unfair Trade Practice through false representation of food quality and appearance. Section 2(11) CPA 2019: Deficiency of Service by both the restaurant and the platform. Food Safety and Standards Act, 2006: Obligations regarding accurate food description and labelling. E-Commerce Rules 2020: Platform obligation to ensure accuracy of product listings. IT Rules 2026: Obligation to implement pre-publication AI-content detection for large platforms. |
Legal analysis | The restaurant is primarily liable for the deceptive listing. The platform is jointly liable for providing the technological infrastructure that enables this deception without any detection safeguard. The platform profits from each order placed, including orders influenced by deceptive AI images. It cannot claim neutrality when it directly benefits from and facilitates the deception. Consumer courts regularly entertain food delivery complaints, making this a practical avenue for relief. |
CASE STUDY 09 The Investment Platform With AI-Generated Financial Advice | |
|---|---|
Platform Type | Online Investment and Financial Advisory Platform |
The Situation | A financial technology platform offers personalised investment advice through an AI chatbot. The AI analyses a user's financial profile and generates specific investment recommendations. The AI model was trained on market data up to a certain date and has not been updated. It recommends a consumer invest a significant portion of their savings in a financial instrument that has since become extremely high-risk due to changed market conditions. The platform presents the AI's recommendations with a confidence-inspiring interface design that resembles a certified financial advisor consultation. The consumer suffers significant financial loss. |
Consumer Harm | Direct financial loss from acting on outdated or incorrect AI investment advice. The consumer may have made different decisions with proper human advice. The financial harm in investment cases can be life-altering, particularly for middle-class consumers investing retirement savings. |
Applicable Laws | Section 2(47) CPA 2019: Unfair Trade Practice through deceptive interface suggesting certified financial advice capability. Section 2(11) CPA 2019: Deficiency of Service. Securities and Exchange Board of India regulations: Requirements for investment advisors to be registered and qualified. DPDP Act 2023: Obligations in processing sensitive financial personal data. Consumer Protection Act Chapter VI: If the AI advisory system is treated as a product, no-fault liability may apply for the defect in its outdated training data. |
|---|---|
Legal analysis | This case illustrates an important gap: SEBI registration requirements for investment advisors were designed for human advisors. AI advisory systems exist in a regulatory grey zone. The platform is liable under consumer protection law for both the deficiency of service and the unfair trade practice of simulating certified advice without being a certified advisor. SEBI must urgently clarify how AI investment platforms are to be classified and regulated. |
CASE STUDY 10 The Online Education Platform With AI-Generated Course Content | |
|---|---|
Platform Type | Online Learning and Certification Platform |
The Situation | A professional education platform sells certification courses in technical and professional subjects. To reduce content creation costs, the platform uses AI to generate course materials, quizzes, and assessments. The AI-generated content contains errors, outdated information, and in some subjects, fundamentally incorrect explanations of core concepts. Students pay significant course fees and receive certificates after completing the AI-generated curriculum. Employers who accept these certificates in good faith find that certificate holders lack the knowledge the certification claims to verify. Students who relied on the certificate for career advancement discover their knowledge is inadequate when they begin work. |
Consumer Harm | Financial loss from course fees paid for unreliable content. Career harm for students whose credentials are discredited. Harm to employers who make hiring decisions based on misleading certifications. The harm is compounded because it affects both the consumer-student and the third parties who rely on the certification. |
Applicable Laws | Section 2(47) CPA 2019: Unfair Trade Practice through false representation of course quality and educational value. Section 2(11) CPA 2019: Deficiency of Service in the delivery of education. Section 84 and 85 CPA 2019: Product liability provisions if AI-generated course content is treated as a defective product. Consumer Protection Act Section 2(10): |
Defect in service for courses containing incorrect professional information. Distance Education Bureau and UGC guidelines on online education quality standards. | |
Legal analysis | The platform is directly liable for the content it sells. Unlike a marketplace where a third-party seller provides defective goods, here the platform is itself the seller of the defective educational content. The use of AI to generate course material is not inherently wrong, but deploying AI-generated content without expert human review in technical professional certification courses is a clear deficiency of service. Students are entitled to refunds. Students who suffered career harm due to the unreliable certification may have additional grounds for compensation. |
Gaps In The Legal Framework
The case studies above show us that existing Indian laws can, in many situations, be applied to provide some remedy to consumers harmed by AI. But the application is incomplete, uncertain, and often practically very difficult. This section identifies the specific gaps in the law that make it hard for consumers to get justice, and explains why each gap matters.
The Legal Gap | What the Law Currently Says | Why This Harms Consumers | What Needs to Change |
|---|---|---|---|
AI is not defined as a product or a service | The Consumer Protection Act covers both products and services but does not mention AI at all. Courts and consumers are left to argue about which category applies. | If AI is a product, no-fault liability applies and consumers get stronger protection. If it is a service, consumers must prove negligence, which is much harder. The uncertainty benefits companies, not consumers. | Parliament must amend the CPA 2019 to expressly define AI systems and assign them to a liability category based on their risk level and the nature of their function. |
Consumer must prove causation in AI cases | Consumer protection law requires consumers to show that the specific AI action was the direct cause of their harm. AI systems are complex black boxes whose decision logic is invisible to consumers. | A consumer rejected by an AI loan system has no way to know why. A consumer harmed by an AI medical recommendation cannot easily access the model's reasoning. Proving causation is practically impossible without access to the AI system itself. | India should adopt a rule similar to the EU AI Liability Directive: if a company refuses to disclose how its AI system works, courts should be allowed to presume the AI caused the harm. This shifts the burden fairly to the company. |
AI Governance Guidelines have no legal force | India's AI Governance Guidelines issued in November 2025 are voluntary. Companies can read them, agree with them, and then | Companies that take AI safety seriously are at a competitive disadvantage against those that cut corners. Consumers get no protection from | The guidelines must be given statutory backing. At minimum, compliance with core safety guidelines must be made a condition for platforms above a |
The Legal Gap | What the Law Currently Says | Why This Harms Consumers | What Needs to Change |
|---|---|---|---|
do nothing. There is no penalty for ignoring them. | companies that simply choose not to follow the guidelines. | certain size to operate in India. | |
No mandatory AI content labelling law before 2026 | Before the February 2026 IT Rules amendment, there was no law requiring platforms to label AI-generated content. Even now, the rules apply only to large platforms classified as SSMIs. | Millions of consumers on smaller platforms and apps still have no legal guarantee that AI-generated images, articles, or profiles will be disclosed as such. They make decisions based on content they believe is real. | The AI content labelling obligation should be extended to all platforms, not just large ones. A tiered compliance system could apply lighter requirements to smaller platforms while maintaining the core disclosure obligation. |
CCPA has not actively targeted AI harms | The Central Consumer Protection Authority has strong powers to investigate AI-driven consumer harm on its own initiative. It has not yet done so in any systematic way. | The absence of enforcement means companies face no consequence for ignoring consumer protection obligations in the AI space. The law exists on paper but not in practice. | The CCPA must establish a dedicated AI and Digital Harms division. It should set up a specialist team, conduct proactive platform audits, and initiate suo motu proceedings against platforms with documented patterns of AI-driven consumer harm. |
Consumer courts struggle with technical AI evidence | District consumer courts are designed for relatively straightforward disputes. AI cases require understanding of machine learning, algorithmic decision-making, and technical audit reports. Most consumer court judges have no | Complex AI cases are often dismissed or resolved without meaningful engagement with the technical facts. Companies with technical resources can overwhelm individual consumers in technical proceedings. | The National Consumer Disputes Redressal Commission should establish a panel of technical experts to assist consumer courts. The CPA should be amended to allow courts to appoint independent AI auditors at company expense |
The Legal Gap | What the Law Currently Says | Why This Harms Consumers | What Needs to Change |
|---|---|---|---|
training in these areas. | when technical evidence is contested. | ||
Sector-specific AI regulations are missing | Health, finance, education, and housing are heavily regulated industries. But none of the sector regulators (SEBI, IRDA, NMC, UGC) have issued clear rules about how AI can be used within their domains. | An AI medical advice app, an AI financial advisor, and an AI education platform all operate in largely unregulated spaces. Consumers in these high-stakes sectors have less protection from AI harm than consumers buying physical goods. | Each major sectoral regulator must issue AI-specific guidelines that define what AI applications are permitted, what human oversight is required, and what disclosure must be made to consumers. |
\ What Needs To Change And What You Can Do
A. What The Government Must Do
1. Pass A Dedicated AI Law
India needs a law specifically designed for AI. This law should define what AI is in legal terms, classify AI systems by the level of risk they pose to consumers, impose mandatory safety requirements on high-risk AI, create a clear liability framework that answers whether AI is a product or a service, and establish penalties significant enough to deter large platform companies.
This law does not need to be hundreds of pages long. The core rules can be concise. What matters is that they are binding, enforceable, and clear.
2. Give The Ccpa A Dedicated AI Division
The Central Consumer Protection Authority must set up a specialised team for AI-related consumer complaints. This team should proactively monitor major platforms, conduct random audits of AI
systems used in consumer-facing applications, respond to consumer complaints about AI-generated deception, and publish transparency reports so that the public knows which platforms are compliant and which are not.
3. Make AI Content Disclosure Universal
The obligation to label AI-generated content must apply to all platforms, not just large ones. A consumer on a smaller shopping site or a regional matrimonial portal deserves the same right to know if they are looking at an AI-generated image as a consumer on a major national platform.
4. Give Sectoral Regulators Clear AI Mandates
SEBI must regulate AI investment advisors. The National Medical Commission must set rules for AI health advice apps. The University Grants Commission must set standards for AI-generated educational content. The Insurance Regulatory and Development Authority must address AI insurance assessment tools. Each of these regulators must publish clear rules within 12 months.
5. Adopt The EU's Rebuttable Presumption Approach
If a company's AI system causes harm and the company refuses to explain how the AI reached its decision, Indian courts should be empowered to presume that the AI caused the harm. This one change would transform the practical ability of ordinary consumers to get justice in AI cases.
B. What Platforms Must Do
1. Implement AI-Image Detection Immediately
Every platform that allows users to upload images must deploy AI-image detection. This technology exists. It is not expensive for large platforms. It can identify AI-generated images with reasonable accuracy. Not deploying it is a choice, and under current law it is an increasingly costly choice.
2. Label All AI-Generated Content Clearly
When content was created or significantly altered by AI, this must be disclosed clearly and prominently to the consumer before the consumer acts on that content. The disclosure must be visible.
It cannot be hidden in terms of service or revealed only on a separate page that consumers are unlikely to read.
3. Maintain Human Review For High-Stakes Decisions
When AI makes a decision that significantly affects a consumer, such as rejecting a loan application, denying an insurance claim, or removing an account, a qualified human being must review that decision before it is final. Consumers must be given a meaningful opportunity to challenge the AI decision and have it reviewed by a person.
C. What You As A Consumer Can Do Right Now
While waiting for the law to catch up, individual consumers are not powerless. Here is what you can do today:
1. Report deceptive AI content to the platform using its official grievance mechanism. Every large platform must have a grievance officer who must respond within 7 days under the current IT Rules. 2. File a complaint with the National Consumer Helpline at 1915. This is free and can be done online or by phone. 3. File an online consumer complaint through the Edaakhil portal at edaakhil.nic.in. You can file cases up to Rs. 50 lakh at the District level, up to Rs. 2 crore at the State level, and above Rs. 2 crore at the National level. 4. If the content is AI-generated and harmful, report it to MeitY through the Grievance Appellate Committee portal. 5. Always do a reverse image search on any product, profile, or property photograph before making a payment. Google Images and other tools can often reveal if a photo was AI-generated or taken from another source. 6. Do not rely solely on AI-generated health, legal, or financial advice. Always verify important advice with a qualified human professional.
As an Indian consumer, you have the right to be protected from hazardous goods and services, the right to be informed about quality and standards, the right to be heard and seek redressal, and the right to consumer education. AI does not remove these rights. The law must ensure that technology serves consumers, not deceives them.