Latest Search
Quote
| Back Zoom + Zoom - | ||||
|
CN Watchdogs Rectify AI Tech Misuse, Handles 3,500+ Violating AI Products
Recommend 20 Positive 28 Negative 18 |
||||
|
||||
|
Since the launch of the “Clear and Bright - Rectification of AI Technology Misuse” special action in April 2025, the Cyberspace Administration of China (CAC) has aggregately handled over 3,500 violating AI products, such as mini-programs, applications, and intelligent agents, cleaning up over 960,000 pieces of illegal information, and handling over 3,700 accounts, achieving positive progress in various tasks, according to the WeChat public account “Cyberspace China”. Regarding key website platforms, in handling violating AI products, TENCENT (00700.HK) has standardized application management, raised entry thresholds, optimized inspection mechanisms, and rejected or handled over 570 violating mini-programs and applications, sources divulged. In cleaning up tutorials and products of violating AI, WB-SW (09898.HK) has used strategy identification, user reports, and other multi-channel reviews to cumulatively handle over 4,800 pieces of incompliant content. In strengthening security management measures, Douyin has established a “red-blue confrontation” mechanism, simulating attack scenarios, fixing potential security loopholes, and optimizing the model's ability to identify false information. Moreover, in implementing content identification requirements, key platforms such as BABA-W (09988.HK)(BABA.US) and KUAISHOU-W (01024.HK) have actively promoted the implementation of metadata implicit identification. In preventing security risks in key areas, RedNote (Xiaohongshu) has input professional domain data during the model's post-training phase to enhance the model's understanding of issues in key areas such as healthcare, finance, and minors. AASTOCKS Financial News Website: www.aastocks.com |
||||
Copied to Clipboard
 |
||||
