The Ministry of Electronics and Information Technology (MeitY) has public consultation on developing guidelines for the governance of artificial intelligence (AI). The guidelines have been issued in a , which is the result of a ministry sub-committee that looked into existing issues and offered recommendations.
Principles, such as transparency, security, accountability etc., have been incorporated into the following proposed guidelines.
Inter-ministerial AI co-ordination committee. This would see various ministries and their departments collaborate with institutions and regulators to ensure governance measures are suited to the realities of existing and possible risks. The committee would have official and non-official industry experts.
Technical secretariat. The secretariat, who would advise and co-ordinate the committee, would work on identifying existing and potential risks, facilitate the development of metrics, and engage with industry experts to foster appropriate and novel solutions.
AI incident database. The technical secretariat would maintain a database of industry problems to build evidence and guide any future response to avoid a similar issue or outcome. Reporting protocols shall be defined for public and private sector entities, as well as public-private partnerships to follow. An incident can include a cyber incident, or cyber security incident etc., already defined under the law. It can also include suspected or potential attack or vulnerability against AI systems, attempt to harm certain individuals/businesses/societies, unexpected emergent behaviour, etc.
Voluntary commitments on transparency. The technical secretariat would engage with the industry encouraging voluntary commitments on transparency. The secretariat would collaborate with industry to reach a consensus on voluntary commitments related to AI’s intended purposes, transparency reports, disclosure of development or testing, or monitoring or security, and other related data etc. These commitments would complement existing legal frameworks.
Examining suitability of measures addressing AI risks. Technology would be used to model the process between datasets, users, applications etc., to identify and track negative outcomes. This would include a dedicated effort to identify and address gaps hindering better accountability and prompt action.
Sub-group under Digital India Act. The group would work with the ministry providing suggestions on legislation to strengthen and harmonise the existing legal framework, regulatory and technical capacity, and adjudicatory setup requirements for grievance redressal systems. This can include online grievance redressal systems in institutions, public and private entities, and public-private partnerships.
All comments on the guidelines must be submitted to the ministry by 27 February 2025.
