NEW DELHI, Aug 13: The Supreme Court on Thursday asked the Centre and other concerned authorities to consider suggestions submitted by a petitioner seeking stronger ethical and transparency standards for artificial intelligence (AI) systems used in surveillance and content moderation.
A bench comprising Chief Justice of India Surya Kant and Justices Joymalya Bagchi and V Mohana was hearing a public interest litigation filed by advocate Narendra Kumar Goswami.
At the outset, the petitioner told the court that he was pursuing a prayer seeking disclosure of all existing and proposed high-risk AI systems deployed by government ministries and agencies for welfare schemes, policing, surveillance and content regulation.
“Tell us what can we do,” the bench asked the petitioner.
Goswami was then asked to forward his writ petition to the government as a representation. The bench observed that the petition covered a comprehensive range of issues.
The court noted that it was not composed of technical experts and that the matters raised largely fell within the policy domain. It observed that the concerns would require appropriate policy or regulatory measures.
The bench also took note of the fact that Goswami had submitted a representation to the government in February this year concerning the use and regulation of AI.
Without expressing any opinion on the merits of the issues raised, the Supreme Court disposed of the petition at this stage and directed the authorities to consider the suggestions and take appropriate steps. It also granted the petitioner liberty to supplement his representation.
The PIL seeks the constitution of a high powered expert committee to formulate binding standards covering AI ethics, algorithmic impact assessments and bias audits.
It also calls for greater transparency in AI-enabled surveillance and content moderation, human oversight of high-risk government AI applications and stronger safeguards for personal data.
Among other demands, the petitioner has sought comprehensive algorithmic impact assessments and independent bias audits of AI systems used in welfare allocation, predictive policing, facial recognition and content moderation.
The plea further seeks public disclosure of an inventory of such systems and calls for parliamentary legislation governing the development and deployment of AI by the State.
The proposed framework, according to the petition, should establish accountability mechanisms while ensuring that the use of advanced technologies by government agencies remains transparent, subject to human oversight and consistent with fundamental rights.