|
Work item:
|
Y.INSA-HAC
|
|
Subject/title:
|
Dynamic human-AI collaboration mechanism for intelligent network status awareness
|
|
Status:
|
Under study
|
|
Approval process:
|
AAP
|
|
Type of work item:
|
Recommendation
|
|
Version:
|
New
|
|
Equivalent number:
|
-
|
|
Timing:
|
2027-11 (Medium priority)
|
|
Liaison:
|
ITU-T SG2, SG20, SG21, ETSI
|
|
Supporting members:
|
China Unicom, China Information Communication Technologies Group, Peng Cheng Laboratory, Beijing University of Posts and Telecommunications, China Mobile Communications Corporation, China Telecommunications Corporation, Institute of Computing Technology Chinese Academy of Sciences
|
|
Summary:
|
With the increasing integration of artificial intelligence into network management, challenges regarding interpretability, adaptability, and trust persist, particularly under non-routine or high-risk conditions where autonomous models may lack sufficient confidence. While existing ITU-T Recommendations (e.g., ITU-T Y.3184) have established foundations for automated network status analysis, mechanisms for dynamic authority arbitration and synergistic interaction between human experts and AI agents remain undefined. Consequently, there is a critical need to establish a standardized mechanism that enables adaptive control switching based on risk and confidence evaluation.
Human-AI Collaboration (HAC) is a dynamic technical mechanism embedded within intelligent network operations. It executes adaptive task allocation and authority switching between AI agents and human experts through deterministic interaction logic and functional interfaces, triggered by real-time conditions such as low confidence levels, unknown anomalies, or high-risk scenarios.
This Recommendation provides an overview of intelligent network status awareness enhanced by dynamic human-AI collaboration, illustrating the integration of human-AI interaction within the network sensing, analyzing, and reporting layers. Building on this foundation, it specifies the core collaboration mechanisms, defining the procedures for dynamic task allocation triggered by risk and confidence assessments, the selection of distinct collaboration modes, and the execution of synergistic actions. Furthermore, it describes the evolutionary closed-loop process that transforms operational feedback into structured knowledge to continuously refine network awareness capabilities.
|
|
Comment:
|
-
|
|
Reference(s):
|
|
|
Historic references:
|
|
Contact(s):
|
|
| ITU-T A.5 justification(s): |
|
|
|
|
First registration in the WP:
2026-03-06 15:02:58
|
|
Last update:
2026-03-06 15:14:13
|
|
