Artificial Intelligence (AI) has transitioned from a vision of the future to a core driver of governance, commerce, security, and social interaction. AI-powered systems are now routinely used in areas such as predictive analysis, facial recognition, healthcare diagnostics, and other automated actions. While these technologies offer efficiency, scalability, and data-driven precision, they also raise serious concerns relating to transparency, accountability, discrimination, and unchecked power. The increasing reliance on algorithmic decision-making has highlighted the risks of bias, AI hallucinations, lack of credibility, and inconsistency, all of which harm vulnerable communities, particularly when AI systems operate without meaningful oversight or remedies.
Globally, these concerns have led to legislative and regulatory responses aimed at governing AI through ethical principles and accountability mechanisms. In India, despite significant public and private sector adoption of AI, a comprehensive statutory framework specifically addressing ethical deployment, algorithmic bias, and transparency has remained absent. It is in this context that the proposed Artificial Intelligence (Ethics and Accountability) Bill, 2025 (the Bill), was presented to Parliament on December 17, 2025, to establish a structured legal framework to regulate AI technologies in sensitive and high-impact domains.
Background of the Bill
The Bill was presented as a private member’s initiative and reflects growing ethical and social concerns with regard to emerging technologies. The Bill aims to establish an Ethics and Accountability Framework governing the use of AI in decision-making, surveillance, and algorithmic systems, with the stated objective of preventing misuse and ensuring fairness, transparency, and accountability. The Bill would come into force on a date to be notified by the Central Government through publication in the Official Gazette. The Bill clarifies that its provisions operate in addition to existing laws and do not derogate from them. This ensures regulatory harmony with other legal regimes, including data protection, sector-specific regulations, and constitutional protections.
Key Definitions
The Bill adopts a broad and inclusive definition of “Artificial Intelligence (AI)”, including computer systems capable of performing tasks that ordinarily require human intelligence, including decision-making, language processing, and visual perception. This expanded definition would ensure that both present and evolving AI technologies fall within the regulatory ambit of the Act.
Additionally, the Bill explicitly recognises “algorithmic bias” as systematic errors leading to unfair outcomes, signalling awareness of one of the most pressing ethical risks associated with AI systems.
Further, the Bill defines “stakeholders” to include developers, deployers, users, and affected individuals or communities, thereby acknowledging that AI governance must account for both creators of technology and those subject to its consequences.
Establishment of the Ethics Committee for Artificial Intelligence
The Bill lays down provisions for the establishment of an Ethics Committee for Artificial Intelligence by the Central Government. The Committee would be envisaged as a multidisciplinary oversight body comprising a Chairperson with expertise in ethics and technology, representatives from academia, industry, civil society, and government, as well as experts in law, data science, and human rights. The Committee would be supported by officers and employees appointed by the Central Government, with their service conditions to be prescribed through rules.
Functions of the Ethics Committee: The Ethics Committee would be assigned both regulatory and developmental duties, including developing and recommending ethical guidelines for AI technologies, monitoring compliance with ethical standards, and reviewing instances of misuse, bias, or violations of the Act. The Committee would also be responsible for promoting awareness and capacity-building among stakeholders, recognising that ethical AI adoption requires education and institutional readiness alongside regulation. Additionally, the Committee may perform such other functions as may be prescribed by the Central Government, allowing the regulatory framework to evolve alongside technological advancements.
Funding of the Committee: The Bill provides for grants from the Central Government to fund the Ethics Committee, with expenditures charged to the Consolidated Fund of India. The Committee would be required to submit an annual report to the Central Government, providing a detailed account of the activities, findings, and recommendations undertaken during the previous financial year. The report would also need to be laid before Parliament within 6 months, ensuring transparency, accountability, and legislative oversight over the functioning of the AI governance framework.
AI-Based Surveillance
The Bill introduces explicit restrictions on the use of AI in surveillance, requiring that such deployment be limited to lawful purposes and subject to prior approval by the Ethics Committee. Further, AI systems used in critical decision-making areas such as law enforcement, financial credit, and employment would be subjected to heightened ethical scrutiny. These systems would be prohibited from discriminating solely on the basis of race, religion, or gender and must undergo stringent ethical review by the Committee. By regulating AI in high-impact domains, the Bill seeks to ensure that automated systems do not undermine constitutional guarantees of equality and fairness.
Obligations of AI Developers
The Bill places significant responsibility on AI developers, recognising their pivotal role in shaping system outcomes. Developers would be required to ensure transparency by disclosing the intended purpose and limitations of AI systems, the data sources and methodologies used for training, and the reasons for decisions made by AI systems that affect individuals. In addition, developers would be required to actively prevent algorithmic bias by conducting regular audits, ensuring diversity and inclusivity in training datasets, and withdrawing AI systems exhibiting significant bias until corrective measures are implemented. The requirement to maintain records of ethical compliance would further strengthen traceability and accountability across the AI lifecycle.
Grievance Redressal Framework
To address complaints arising from AI deployment, the Bill establishes a grievance redressal mechanism that would enable affected individuals or groups to file complaints with the Ethics Committee. The Committee would be empowered to investigate such complaints and recommend penalties or remedial measures as prescribed. This mechanism introduces procedural safeguards and provides a formal avenue for redress, addressing a critical gap in existing AI governance structures, where affected individuals often lack effective remedies.
Penalties for Non-Compliance
The Bill provides that any non-compliance may attract financial penalties extending up to INR 5 crores, depending on the severity of the offence. It also states that non-compliance may also lead to suspension or revocation of licences for deploying AI systems. In cases of repeat violations, the Bill contemplates criminal liability, reflecting a strong deterrent approach.
Conclusion
The proposed Artificial Intelligence (Ethics and Accountability) Bill, 2025, is an attempt to establish a comprehensive legislation to address the ethical, social, and legal challenges posed by AI technologies in India. By combining institutional oversight, accountability of developers, restrictions on high-risk applications, and penalties, the Bill seeks to balance innovation with responsibility. The proposed Bill reflects the growing recognition of the need for AI governance to ensure transparency, fairness, and democratic values, ensuring that technological progress serves the broader public interest rather than undermining it.
Authors: Manisha Singh and Shivi Gupta
First Published by: Mondaq here