Indian Advisory Requires Approval for “Unreliable” AI Tools
The Indian government has issued an advisory requiring companies to obtain government approval before releasing certain artificial intelligence (AI) tools.
This advisory, issued by the Ministry of Electronics and Information Technology (MeitY), focuses on “unreliable” or un-tested AI models, particularly those used by social media platforms.
The primary concern of the government is potential harm to citizens and democracy through the spread of misinformation and biased content. Instances of this, the law states, have been identified in recent times with experimental models from various AI platforms.
To address these concerns, MeitY mandates approval for “unreliable” AI models, which the law defines as those still under development or with the potential for inaccurate outputs. Additionally, platforms utilizing such models, including popular names like Google’s Gemini, ChatGPT, and Krutrim AI, will be required to:
- Label themselves as “under testing”: This transparency measure aims to inform users about the potential limitations and unreliability of the AI model.
- Seek user consent: Platforms must explicitly obtain user consent, clearly stating that the employed AI model might generate inaccurate or erroneous information.
- Embed metadata: Digital publishing platforms must develop a system to embed metadata or unique identification codes within synthetically generated content. This would facilitate tracing the origin of such information.
The law has received mixed reactions. While the government defends the move as an essential safeguard for public safety, some startups believe it stifles innovation.
Companies failing to comply with the law’s provisions risk facing penal actions under the Information Technology Act and other relevant legal frameworks.
As India navigates the rapidly evolving landscape of AI, this advisory serves as a reminder of the ongoing discussion about balancing technological advancement with responsible development and safeguarding public well-being.
Whether the proposed measures prove effective in achieving this balance remains to be seen, but the controversy it has sparked highlights the complexities involved in regulating AI in today’s world.
