India has decided to backtrack on a recent AI advisory that mandated tech companies to obtain government approval before releasing new AI models.
The Ministry of Electronics and IT released an updated AI advisory on Friday after consulting with industry stakeholders. This revised advisory no longer requires companies to seek government approval before launching or deploying an AI model in the South Asian market. Instead, firms are encouraged to label under-tested and unreliable AI models to make users aware of their potential shortcomings.
The revision comes after India’s IT ministry faced criticism earlier this month from various prominent figures. Less than a year ago, the ministry had opted not to regulate AI growth, citing its importance to India’s strategic interests.
Although the advisory is not legally binding, the ministry stated earlier this month that it indicates the “future of regulation” and expects compliance from entities.
According to TechCrunch, the advisory underscores that AI models must not facilitate the sharing of unlawful content under Indian law and should prevent bias, discrimination, or threats to the integrity of the electoral process. Intermediaries are encouraged to employ “consent popups” or similar methods to explicitly notify users about the potential unreliability of AI-generated output.
The ministry maintains its focus on ensuring the easy identification of deepfakes and misinformation. Intermediaries are advised to label or embed content with distinct metadata or identifiers. The requirement for firms to develop a method to identify the “originator” of specific messages has been removed.
Source: Social Samosa
