The Ministry of Electronics and Information Technology (MeitY) issued a new advisory on Artificial Intelligence (“AI”) on March 15, 2024 (“New Advisory”), scrapping the earlier version issued on March 1, 2024. The New Advisory has diluted the strict position taken by MeiTY in its earlier advisory.
1. CHANGES
Though, the New Advisory has retained majorly all clauses of the earlier version, there are few changes brought out in terms of the regulatory framework. These changes have been summarized below:
A. No More Prior Government Approval
Previously, intermediaries and platforms were required to ensure that a proper government approval was in place prior to making under-testing/ unreliable AI tools available to Indian public.
The New Advisory scraps the previous approval mechanism which was loaded with ambiguities. Specifically, there was a lack of clarity on definitions for “unreliable” models and the process for obtaining approval.
B. Focus on Labelling
The New Advisory underscores the significance of user awareness regarding AI-generated content. Intermediaries and platforms are required to ensure appropriate labelling of AI, particularly which is under-tested/ unreliable, regarding the fallibility of outputs.
Intermediaries and platforms have to ensure user awareness regarding fallibility of AI tools/systems and their outputs. User awareness should be created using consent popup or any other relevant mechanism.
This transparency empowers users to critically evaluate the information they encounter online and mitigates the potential for manipulation through AI-powered misinformation campaigns. AI-generated content, especially those vulnerable to misuse like deepfakes, must be clearly labelled for user awareness.
C. No Reporting
Under the previous advisory intermediaries and platforms were also mandated to submit an Action Taken-cum-Status Report (“ATS Report”) to MeitY within a 15-day timeframe. The New Advisory has removed the reporting requirement.
2. ISSUES
Even though the New Advisory states that it is applicable to Platforms, it does not clarify which entities are covered by the expression ‘Platforms’ leaving a question mark on scope of its applicability to non-intermediaries using or making available AI tools.
Further, the is no clear guidance about testing of AI tools/systems. The government has not clarified any standards for testing the AI tools. Intermediaries and platforms have been asked to ensure labelling of under-tested/ unreliable AI tools without any clear prescription on standards of testing AI tools or clarity about authority which will certify testing and reliability of AI tools.
3. IMPACT
The changes brought out in the New Advisory mark a shift in India’s regulatory approach; from zero regulation to a more measured stance on AI regulation. The New Advisory has addressed the concern raised about implementation of prior approval mechanism by removing that requirement.
Despite the removal of approval mechanism, intermediaries and platforms still face significant risk due to lack of clarity and ambiguity in standards of testing unreliable AI systems. While the New Advisory might reduce compliance costs for intermediaries and platforms, it might inadvertently increase their chances of losing legal immunity (safe harbor) under Indian law.
As of now, intermediaries and platforms are required ensure compliances which were enlisted in the earlier advisory except seeking government approval for using under-testing/unreliable AI tools and submitting ATS Report. We have summarized the compliances mentioned under the advisory issued on March 1, 2024, you can read it here: https://shorturl.at/JKQ13