Context:
The US Bureau of Industry and Security BIS introduced a tiered framework for controlling exports reexports and incountry transfers of AIrelated technology. The aim is to align with national security and foreign policy interests and to foster a secure and trusted technology ecosystem for AI.
Scope of Technology Regulated
Target specifics include advanced computing chips and closed AI model weights in particular.
AI models are simply software programs performing mathematical calculations in producing outputs such as analysis or even media. The pattern of these operations, called the architecture of the model, determines its performance.
Proposed Framework
- Tier 1
- Unrestricted access for 18 US allies including Australia, Canada, South Korea and the UK.
- Tier 2
- Access is denied with volume caps and needs authorization specifically for chips destined for highly sophisticated AI models.
- China and India fall into this tier and require VEU authorization for sensitive transactions.
- Tier 3
- No access for arms embargoed countries like North Korea Iran and Russia.
Rationale for Restrictive Measures
- Prevent technology from reaching countries of concern or adversaries
- Address risks such as:
- Enhanced military capabilities of adversaries.
- Development of weapons of mass destruction.
- Offensive cyber operations.
- Human rights abuses like mass surveillance.
- Reduce the risk of Chinese companies using foreign subsidiaries to circumvent controls.
- Address risks such as:
Impact on India
- Indian data centers will need VEU clearance to process the advanced AI chips.
- Indian companies with VEU clearance will be allowed to use the technology for civilian and military purposes but not for nuclear-related applications.
- Experts say India does not fit in the Tier 1 category of trusted allies, possibly due to chip leakages to Russia.
