Unverified50% confidenceFactExact time
The Decoder reported on October 31 that an email sent to members of Trusted Access in the Daybreak Blue cybersecurity program stated that, starting November 29, certain security drawing data would be used for model training even if users had turned off the ordinary "improve the model for everyone" toggle.
1
Sources
50%
Confidence
Short-term (~14 days)
Relevance
11/1/2026
First Seen
Valid until: 11/15/2026
Sources
DeepSeek Open-Sources SuperAgent: Agents Must Learn to Take On Work Continuously
bilibili镇渊AI11/1/2026
Related Entities
Related Claims
UnverifiedMainstream large models all undergo safety alignment training before release, causing them to refuse requests such as generating malware code, assisting with cyberattacks, or producing illegal content.64% similarUnverifiedWhen uploading operational data to a third-party AI platform, one should confirm whether the data is used for model training, the method of storage encryption, deletion rights, and compliance certifications such as SOC 2.64% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1124724API
curl https://kongchang.com/api/v1/knowledge/claims/1124724MCP
get_claim(id=1124724)