India AI Ethics & Accountability Bill 2025: Status & Provisions

Bill Details

The Artificial Intelligence Ethics and Accountability Bill 2025 is a Private Member's Bill, meaning it was introduced by an individual MP rather than the ruling government.^{SCC Online} This distinction matters: Private Members' Bills rarely become law, but they signal emerging legislative priorities and often influence government-sponsored legislation.

The bill proposes four major mechanisms: a statutory Ethics Committee, mandatory ethical reviews for high-risk AI, bias audit requirements, and financial penalties up to INR 5 crore (~$590,000 USD).^Tsaaro

Parliamentary Process

For any Private Member's Bill to become law in India, it must clear six stages. The bill has completed only the first.^{Lok Sabha}

The Speaker may refer the bill to a standing or select committee for detailed examination. If referred, the committee reviews the bill, takes expert testimony, and publishes recommendations. This process typically takes months. After that, the bill must be scheduled for debate in the Lok Sabha, where Private Members' Bills compete for limited floor time.

The bill is notable because it represents the first formal attempt by any Indian legislator to create AI-specific statutory obligations.^{SCC Online} Whether this particular bill advances or not, the provisions it proposes indicate where parliamentary thinking is heading.

Key Provisions (Proposed)

How It Fits India's Existing Framework

India does not lack AI governance. It lacks AI-specific legislation with enforcement teeth. The bill would fill a specific gap: statutory requirements for high-risk AI use cases that currently sit in a regulatory gray zone.^{PIB 2025}

MeitY's AI Governance Guidelines (November 2025) are voluntary. They establish principles and governance bodies but carry no penalties for non-compliance.^{MeitY PDF}

The DPDPA (August 2023) covers data protection broadly, including AI training data and automated processing. It has real penalties (up to INR 250 crore) but was not designed specifically for AI governance.

Sector regulators (RBI, SEBI, ICMR, IRDAI, CERT-In) each address AI within their domains, but coverage is uneven. Full sector regulator breakdown here.

ISO 42001, adopted as an Indian standard by BIS, provides a certifiable management system but is voluntary. Organizations pursuing certification can also align with the NIST AI Risk Management Framework, which shares many of the same risk-based principles.

Across all these instruments, data governance remains a recurring theme: every framework addresses how AI training data is collected, processed, and retained, yet none mandate a pre-deployment ethical review of the algorithm itself. The proposed bill would create that requirement, using employment screening as a primary example.

Comparison with EU AI Act Enforcement

The bill's approach shares structural similarities with the EU AI Act but differs in scale and scope.^{European Commission National Law Review}

The EU AI Act's penalties are orders of magnitude larger and apply across a broader range of AI systems. The Indian bill focuses specifically on use cases with direct impact on individual rights, which reflects India's broader governance philosophy of targeting harm rather than regulating technology broadly. For a detailed side-by-side analysis, see our India vs EU AI Act comparison.^{National Law Review}

What to Prepare Regardless

Whether this bill passes or not, the direction is clear. India is moving toward more structured AI governance. Organizations operating in the Indian market should prepare for eventual regulatory requirements.

• 01
  Document your AI systems. Maintain an inventory of all AI systems, their purpose, data sources, and risk profile. Already recommended under MeitY's guidelines and required under ISO 42001.
• 02
  Conduct voluntary bias audits. Especially for AI in hiring, performance evaluation, customer segmentation, or any process affecting individuals based on protected characteristics. The AI glossary covers algorithmic bias and fairness metrics. The ISO 42001 resource center includes control mappings for bias testing under Annex A.
• 03
  Establish an internal ethics review process. A documented review process for high-risk AI deployments demonstrates responsible governance and positions your organization ahead of regulation. Download governance templates to accelerate your documentation.
• 04
  Map your AI systems to the bill's categories. Identify which systems would fall under surveillance, law enforcement, or employment categories. Run a gap analysis against the proposed requirements. Professional credentials such as the IAPP AIGP certification or relevant IT certifications can help your team build the skills needed for these assessments.
• 05
  Track the parliamentary calendar. Private Members' Bills can be revived in subsequent sessions. Monitor Lok Sabha proceedings for any government-sponsored AI legislation incorporating similar provisions.

Organizations already complying with MeitY's voluntary guidelines or pursuing ISO 42001 certification will find that much of the groundwork is already done. The bill's proposed requirements overlap significantly with existing best practices. As demand for these skills grows, AI governance careers are expanding rapidly across India (see current salary benchmarks), and familiarity with concepts like explainability, risk tiering, and conformity assessment will be essential.

Stay Updated

This page tracks the bill's progress through Parliament. As the bill moves through committee referral, debate, or any amendments, we will update the status and analysis.