In-Short
- India mandates government approval for AI technologies before public release.
- AI developers must disclose potential unreliability in their technologies.
- Deepfakes to be labelled with unique identifiers to prevent misuse.
- AI models must avoid bias and not compromise electoral integrity.
Summary of India’s New AI Advisory
The Indian Ministry of Electronics and Information Technology (MeitY) has issued a new advisory that requires all AI technologies under development to obtain government permission prior to public deployment. This move aims to regulate the burgeoning AI sector and address potential risks associated with AI systems.
Under the new guidelines, developers are also obligated to label their AI outputs, indicating any possible errors or fallibility. A “consent popup” feature is to be implemented to alert users about these potential defects. Additionally, the advisory calls for the labelling of deepfakes with unique metadata to prevent their misuse.
Platforms hosting AI models, including large language models (LLM), are instructed to ensure their products are free from bias and do not threaten the integrity of electoral processes. The advisory has been met with criticism from some industry experts, who argue that it could hinder innovation and progress in the AI field.
Developers are expected to comply with the advisory within 15 days of its release. Although the advisory is not legally binding at the moment, it reflects the government’s intentions and may foreshadow future legislation. IT Minister Rajeev Chandrasekhar emphasized the importance of AI platforms taking full responsibility for their products, even during the testing phase.
Further Information
For more detailed insights on India’s new AI regulations and their implications, please refer to the original article.
