India to Prepare AI Regulation Framework, With Focus on Deepfakes, Safety and Human Welfare

NEW DELHI, Oct 9: India is preparing to take a significant step towards regulating artificial intelligence (AI), with the government planning to release a consultation paper within a month to establish the foundation for a future regulatory framework. The proposed exercise will focus on AI safety, deepfakes, cybersecurity, human welfare, skill development and measures to prevent harmful uses of the technology.

Union Electronics and Information Technology Minister Ashwini Vaishnaw announced the plan on October 8, saying the time had come for India to develop a more comprehensive approach to AI governance. The proposed document is expected to invite views on how the technology should be developed, deployed and monitored while ensuring that its benefits reach different sections of society.

The announcement comes as artificial intelligence becomes increasingly integrated into communication, education, healthcare, financial services, manufacturing and public administration. While the technology offers opportunities to improve productivity and expand access to services, its rapid development has also raised concerns about fabricated content, digital fraud, privacy, cybersecurity and the consequences of increasingly autonomous systems.

The consultation paper is intended to help shape the country’s future regulatory direction. Its announcement does not, by itself, establish a new set of binding AI rules.

Deepfakes emerge as a major concern

One of the government’s principal concerns is the growing sophistication of deepfakes, which use AI to generate or manipulate images, audio and videos to make fabricated material appear authentic.

Such content can imitate a person’s face, voice or mannerisms, making it increasingly difficult for ordinary users to distinguish genuine recordings from synthetic material. The technology can be used for entertainment and legitimate creative work, but it can also facilitate impersonation, financial scams, harassment and the spread of misleading information.

Vaishnaw has described deepfakes as a serious nuisance capable of creating significant problems in society. He has also highlighted the improving quality of AI-generated material, which makes detection and verification more challenging.

The risks extend beyond social media. Fraudsters may use synthetic voices to impersonate relatives, company executives or public officials. Manipulated videos can damage reputations, while fabricated statements attributed to public figures may spread rapidly before they can be verified.

For businesses, deepfakes can create additional security risks. An employee receiving an apparently genuine video or voice instruction from a senior executive may be persuaded to transfer funds or disclose confidential information.

Addressing these threats requires more than identifying manipulated files after publication. Possible approaches include clearer disclosure of AI-generated content, stronger verification procedures, accessible reporting mechanisms and improved tools for identifying synthetic media.

The consultation process could help establish how responsibility should be shared among AI developers, digital platforms, content publishers and individuals who deliberately misuse these systems.

However, any future rules will need to distinguish harmful deception from legitimate uses of AI, including filmmaking, satire, accessibility tools and other creative applications.

Five areas to shape the proposed framework

The government’s planned approach is expected to address five broad areas: safety, humanity, skilling, inclusivity and harm mitigation.

Safety will involve examining the risks associated with AI systems, including the possibility of unreliable outputs, security weaknesses and harmful automated decisions. Depending on how the framework develops, this could involve clearer expectations for developers and organisations deploying AI in sensitive settings.

The human-centred principle is intended to keep people’s interests at the centre of technological progress. AI may assist with decisions, analysis and service delivery, but questions remain about accountability when automated systems produce inaccurate or damaging results.

Skilling is another important consideration because the spread of AI is changing the skills required in many professions. Workers may need training to use AI tools effectively, understand their limitations and adapt to changes in existing job roles.

Inclusivity concerns the distribution of AI’s benefits. The government has emphasised that the technology should not remain accessible only to large corporations or people with advanced technical knowledge.

Harm mitigation will require attention to the potential misuse of AI, including deepfakes, cybercrime, discriminatory outcomes and other forms of digital abuse.

These priorities indicate that the government’s approach is intended to consider both technological risks and the wider social and economic consequences of AI adoption.

Consultation process to help determine future rules

A consultation paper allows policymakers to seek feedback before finalising a regulatory approach. The process can bring together technology companies, researchers, startups, civil society organisations, legal experts and members of the public.

Their input can help identify where existing laws and enforcement mechanisms may be sufficient and where additional safeguards might be needed.

For example, AI developers may seek clarity on their responsibilities when a model produces harmful content. Digital platforms may require guidance on detecting synthetic media, while smaller businesses may need practical compliance standards that do not impose disproportionate costs.

Researchers and civil society groups may focus on transparency, discrimination, privacy and the ability of individuals to challenge decisions influenced by automated systems.

A consultation can also help the government understand differences between low-risk and high-risk applications. An AI tool used to organise personal notes presents different concerns from a system involved in medical recommendations, financial decisions or critical infrastructure.

A proportionate framework could take these differences into account instead of applying identical obligations to every use of the technology.

The precise legal requirements, enforcement arrangements and compliance deadlines will depend on the proposals developed following the consultation process. These details have not been established by the announcement alone.

AI adoption across key sectors

The government’s regulatory plans are developing alongside efforts to expand the practical use of AI throughout the Indian economy.

In healthcare, AI tools can assist with analysing medical images, organising patient information and supporting clinical workflows. Such applications may improve efficiency, but they require safeguards to protect sensitive information and ensure that inaccurate outputs do not compromise patient care.

In agriculture, AI-powered systems can help farmers analyse weather patterns, monitor crops and identify potential pest or disease problems. Digital advisory services could also help make information more accessible to rural communities.

Educational institutions can use AI to support personalised learning, explain difficult concepts and assist teachers with administrative work. At the same time, schools and colleges must address concerns about unreliable information, plagiarism and excessive dependence on automated tools.

Businesses can apply AI to customer support, document processing, inventory management and data analysis. Smaller enterprises may benefit from lower operating costs and improved access to digital services, provided they have the necessary infrastructure and training.

These applications demonstrate why policymakers face a dual responsibility: encouraging productive uses of AI while ensuring that people are protected from foreseeable risks.

Digital inclusion and the skills challenge

India’s large population and diverse economic conditions make digital inclusion a central issue in the country’s technology strategy.

Access to AI tools may vary significantly between urban and rural areas, large companies and small businesses, and individuals with different levels of digital literacy.

If advanced systems remain concentrated among well-funded organisations, the benefits of automation could be unevenly distributed. Smaller businesses may struggle to afford computing resources, while workers without relevant training could find it difficult to adapt to changing employment requirements.

The government’s emphasis on skilling and inclusivity suggests that AI policy will need to consider these differences alongside technical safety.

Training programmes could help employees understand how to use AI applications, verify generated information and protect confidential data. Educational institutions may also need to update curricula to reflect the growing importance of digital literacy, data analysis and responsible technology use.

At the same time, access to AI should not depend solely on formal technical qualifications. User-friendly applications and affordable digital infrastructure can help individuals and small enterprises benefit from tools that were previously available mainly to specialists.

Ensuring broad participation will be important if AI is to contribute to productivity growth across different sectors rather than deepen existing digital inequalities.

Balancing regulation with innovation

A central challenge for policymakers will be to introduce meaningful safeguards without making it unnecessarily difficult to develop useful AI applications.

Excessively complicated rules could increase compliance costs, particularly for startups and smaller developers. On the other hand, weak safeguards could leave users exposed to fraud, privacy violations and other forms of harm.

The consultation process offers an opportunity to examine this balance before detailed requirements are introduced.

One possible consideration is whether obligations should depend on the level of risk associated with an AI application. Systems used in sensitive areas may require stronger testing, documentation and human oversight than tools designed for routine, low-impact tasks.

Another issue is accountability. When AI-generated content causes harm, regulators may need to determine the respective responsibilities of developers, organisations deploying the system and people who misuse its output.

Transparency will also be important. Users may need clear information about when they are interacting with AI, what limitations a system has and how they can seek human assistance when necessary.

The government will have to consider how any future AI-specific rules interact with existing legal provisions governing data protection, online platforms, cybersecurity and consumer rights.

Clear guidance would help businesses understand their obligations while allowing regulators to respond consistently to violations.

International developments add urgency

Governments around the world are examining ways to govern AI while remaining competitive in research and commercial deployment. Their approaches differ, reflecting different legal systems, economic priorities and views about the appropriate role of government.

India’s consultation initiative comes amid growing international concern about synthetic media, automated decision-making and the security of advanced AI systems.

For Indian companies operating internationally, regulatory developments in multiple markets can create additional complexity. Businesses may need to meet different standards for transparency, data handling and risk management depending on where their services are offered.

A clear domestic framework could help organisations plan their investments and compliance strategies while providing a common basis for discussions with international partners.

However, the effectiveness of any framework will depend on its implementation, the resources available to enforcement agencies and the ability of rules to keep pace with rapidly changing technology.

AI systems can evolve faster than traditional policy processes. Regulators will therefore need mechanisms to review emerging risks and update guidance when new capabilities or forms of misuse appear.