The Union Ministry of Electronics and Information Technology (MeitY) has issued a new advisory to platforms and intermediaries regarding the deployment of Artificial Intelligence (AI) models in India.
Platforms or intermediaries must obtain explicit permission from the Centre before launching AI models in India. MeitY previously issued an advisory in December last year to social media platforms, directing them to adhere to existing IT rules, specifically regarding the handling of deepfakes.
Key Points of the Advisory before launching AI models in India
- Permission Requirement: Platforms must seek explicit permission before deploying AI models, ensuring users are informed about their fallibility.
- Due Diligence: Intermediaries must fulfill their obligations outlined in the Information Technology (Intermediary Guidelines and Digital Media Ethics Code) Rules, 2021.
- Content Regulation: Platforms must prevent the hosting, sharing, or modification of unlawful content and avoid bias or discrimination, safeguarding the electoral process.
- User Awareness: Users must be informed about the potential unreliability of AI-generated content through a consent popup mechanism, including account suspension and legal action.
- Labeling Misinformation: Intermediaries facilitating the creation of potentially misleading content must label it with unique identifiers.
Compliance and Consequences
Intermediaries are instructed to prevent the hosting or sharing of unlawful content and to avoid bias or discrimination in their AI models.
Non-compliance could lead to penal consequences, including prosecution under IT Act and other relevant laws. Intermediaries are required to report their compliance within 15 days of receiving the advisory.
In conclusion, the Ministry’s advisory aims to ensure responsible use of AI models by platforms and intermediaries, with a focus on transparency, legality, and accountability.
Announcing the updates, the Union Ministry of Electronics and Information Technology (MeitY) stated:
The use of under-testing/unreliable Artificial Intelligence model(s) and its availability to the users on the Indian Internet must be done so with explicit permission of the Government of India and be deployed only after appropriately labeling the possible and inherent fallibility or unreliability of the output generated.