Navigate the complex regulatory landscape for self-driving vehicles, drones, and robotics including safety standards, liability frameworks, and emerging governance approaches.
Self-driving vehicles represent one of the most complex AI governance challenges, combining safety-critical decision-making with significant liability questions.
| Level | Name | Description | Driver Role |
|---|---|---|---|
| 0 | No Automation | Human performs all driving tasks | Full control |
| 1 | Driver Assistance | System assists with steering OR acceleration | Primary control |
| 2 | Partial Automation | System controls steering AND acceleration | Must monitor |
| 3 | Conditional Automation | System handles most driving, human backup | Fallback ready |
| 4 | High Automation | System handles all driving in defined conditions | Not required in ODD |
| 5 | Full Automation | System handles all driving in all conditions | None |
| Jurisdiction | Framework | Key Requirements |
|---|---|---|
| US (NHTSA) | FMVSS + AV Guidelines | Safety standards, voluntary guidance, exemption process |
| EU | UN ECE Regulations | ALKS (Level 3), type approval requirements |
| UK | Automated Vehicles Act 2024 | Self-driving authorization scheme |
| China | MIIT Regulations | Testing permits, data localization |
Unmanned Aircraft Systems (UAS) are subject to aviation safety regulations with specific rules for AI-enabled autonomous operations.
| Jurisdiction | Authority | Key Rules |
|---|---|---|
| US | FAA | Part 107, Remote ID, BVLOS waivers |
| EU | EASA | EU Drone Regulations 2019/947, 2019/945 |
| India | DGCA | Drone Rules 2021, Digital Sky platform |
| UK | CAA | UK Drone Code, registration requirements |
Beyond Visual Line of Sight (BVLOS) operations typically require AI systems for detect-and-avoid capability. Regulators are developing frameworks for autonomous BVLOS flights including detect-and-avoid standards and remote identification requirements.
Industrial robots and AI systems are subject to machinery safety regulations with emerging frameworks for collaborative robots and autonomous systems.
The new EU Machinery Regulation (2023/1230) includes specific provisions for AI-enabled machinery:
Autonomous systems challenge traditional liability frameworks, leading to new legislative approaches for allocating responsibility.
| Approach | Description | Jurisdiction Examples |
|---|---|---|
| Product Liability | Manufacturer liable for defective products | US (strict liability), EU PLD |
| Operator Liability | Vehicle/system operator bears responsibility | Traditional vehicle insurance |
| No-Fault Insurance | Insurance covers regardless of fault | UK AV Act proposals |
| AI Liability | Specific rules for AI-caused harm | EU AI Liability Directive proposal |
The proposed EU AI Liability Directive would introduce presumption of causality for AI systems and disclosure requirements, making it easier for victims to prove AI-caused harm. High-risk AI systems face stricter liability exposure.