Bureau of Indian Standards: Indian Standards on artificial intelligence (LITD 30 and IS/ISO/IEC 42001)

The Bureau of Indian Standards is India's national standards body. Its website organises Indian Standards by technical departments and sectional committees.⁸ For artificial intelligence the relevant committee is LITD 30, which, according to the India AI Governance Guidelines, develops standards in the area of AI for India and also contributes to the development of international standards, for example in ISO/IEC JTC 1/SC 42 "Artificial intelligence".¹

BIS's approach to AI so far is largely one of adoption: its AI standards are Indian editions of ISO/IEC publications, which is why the IS numbers carry the ISO/IEC edition year (for example IS/ISO/IEC 42001:2023). The government list does not state the dates on which the Indian editions were adopted, and BIS's online catalogue could not be searched for this entry.¹

BIS standards are voluntary by default. They become binding only when a regulation, licence condition or procurement rule requires conformity, a point MediaNama made in December 2025 in describing the role of BIS in the Government's AI governance framework.³

Annexure 6 of the India AI Governance Guidelines lists 26 Indian Standards "published/under development by BIS". Grouped by subject, they are:¹

Five standards are listed as under development: reliability assessment of AI systems; implementation guidance on de-identification of data used in machine learning; verification and validation analysis of AI systems; an overview of differentiated benchmarking of AI system quality characteristics; and guidance for output data quality of generative AI applications.¹

The standard most often discussed in connection with AI governance is IS/ISO/IEC 42001:2023, titled "Information technology - Artificial intelligence - Management system" in the Government's list.¹ As a management-system standard, it is aimed at organisations that develop or use AI and want to set up, run and improve a documented process for managing AI risks and responsibilities, in the same way that organisations adopt ISO management-system standards for quality or information security. Together with IS/ISO/IEC 23894:2023 on risk management and IS/ISO/IEC 38507:2022 on governance, it gives a standards-based route to the risk-based approach to governance that the Government's guidelines favour.¹

Three limits should be noted. First, the sources reviewed do not show when BIS adopted the standard or whether BIS runs a certification scheme for conformity to it. Second, the standard is voluntary, and nothing in the sources reviewed shows a regulator, such as the RBI or SEBI, requiring it. Third, the Government's own guidelines describe standards bodies' work as ongoing, with BIS and TEC expected to develop standards for AI risk taxonomies and certification and to standardise testing, assessment, evaluation and validation procedures.¹

The India AI Governance Guidelines present standards as one of the main non-statutory tools of governance. Their institutional table lists BIS and the Telecommunication Engineering Centre as the standards bodies, with functions that include "developing standards in relation to AI risk taxonomies, certification standards", engagement with global standard-setting bodies and standardising testing and evaluation procedures. Sectoral regulators such as the RBI, SEBI, IRDAI, TRAI, ICMR and the National Health Authority are listed separately, with responsibility for sector-specific rules and enforcement.¹ In the guidelines' list of voluntary instruments, the TEC's fairness standard and the BIS list are cited together as the technical standards available to industry.¹

The approach connects several other Pacta entries. TRAI's July 2023 recommendations proposed that a future AI authority coordinate with TEC on accrediting laboratories to test AI products, and said sector-specific accreditation frameworks would also be needed.⁷ TEC's own fairness standard, TEC 57050:2023, provides a certification-style fairness rating developed by TEC itself, while the BIS list consists of Indian editions of ISO/IEC publications.⁶

Data privacy. In November 2020 BIS established IS 17428, "Data Privacy Assurance", a voluntary standard under which a business can set up a data privacy management system covering lawful and transparent processing, purpose and storage limitation, security safeguards, privacy impact assessments, the evaluation of data processors and staff accountability. MediaNama noted at the time that no regulatory mandate backed it. It covers personal data only and is not AI-specific; it predates the Digital Personal Data Protection Act, 2023.⁴

Data-centre infrastructure. On 5 August 2026 the Minister of State for Electronics and IT, Jitin Prasada, told the Lok Sabha in a written reply on the environmental sustainability of AI data centres that BIS standards published through its LITD 31 committee on cloud computing, IT and data centres address power usage effectiveness, carbon usage effectiveness, cooling efficiency ratio and water usage effectiveness. MediaNama reported that the reply gave no standard numbers or dates and did not say whether compliance is mandatory.⁵

As of 15 October 2026, the most recent verified source on BIS's AI standards is the November 2025 list in the India AI Governance Guidelines, which records 26 published standards and five in development.¹ The sources reviewed for this entry did not allow BIS's own catalogue to be searched, so later additions, the adoption dates of individual standards and any conformity-assessment scheme could not be checked. Readers needing the exact status of a standard should consult BIS's published standards portal.⁸

No evidence was found that any Indian regulator has made conformity to the BIS AI standards mandatory, and the Government's framework continues to treat standards as a voluntary complement to existing law and sectoral regulation.¹²