|
Work item:
|
X.sg-eAI
|
|
Subject/title:
|
Security requirements and guidelines for embodied artificial intelligence systems
|
|
Status:
|
|
|
Approval process:
|
AAP
|
|
Type of work item:
|
Recommendation
|
|
Version:
|
-
|
|
Equivalent number:
|
-
|
|
Timing:
|
-
|
|
Liaison:
|
ITU-T SG21
|
|
Supporting members:
|
China Unicom, ETRI, ZTE Corporation, MIIT China, Ant Group
|
|
Summary:
|
Embodied artificial intelligence systems deeply integrate perception, decision-planning, and physical execution capabilities, enabling intelligent agents to actively comprehend environments, dynamically adapt behaviors, and perform real-world tasks. By bridging virtual intelligence and the physical world through multimodal interaction, embodied artificial intelligence empowers robots and autonomous devices to achieve continuous adaptation and evolution in complex scenarios. However, this deep integration into the physical world introduces operational failures, data breaches, and decision biases as critical security threats. These security threats may directly cause physical damage or personal injury, demanding urgent systemic safeguards.
Therefore, it is necessary to analyze the security threats and security requirements for the embodied artificial intelligence systems to ensure the security of embodied artificial intelligence systems.
This recommendation analyzes the security threats and security requirements for the embodied artificial intelligence systems and provides the security guidelines for it.
|
|
Comment:
|
-
|
|
Reference(s):
|
|
|
Historic references:
|
|
Contact(s):
|
|
| ITU-T A.5 justification(s): |
|
|
|
|
First registration in the WP:
2025-09-22 12:05:27
|
|
Last update:
2025-09-22 12:14:05
|
|
