Unverified50% confidenceDecision RuleExact time
Data passing through machine learning endpoints must be encrypted end to end (TLS 1.3 in transit, AES-256 at rest), and enterprise agreements should guarantee that vendor models neither retain nor train on user data.
1
Sources
50%
Confidence
Long-term
Relevance
10/19/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedTo enable fast retrieval and real-time comparison, many identity verification companies often employ only transport-layer encryption (TLS) while storing data in plaintext or with weak encryption on the server side.75% similarUnverified用户输入通过HTTPS加密传输至服务器，数据是否被纳入模型微调训练集取决于服务商的隐私政策74% similarUnverified子代理提示词加密可采用传输层加密（TLS/mTLS）、应用层加密（AES-256）或令牌化混淆等方案73% similarUnverified传统加密只能保护静态数据与传输中的数据，而AI推理与训练必须将数据加载至内存运算，这个使用中环节是最脆弱的攻击面73% similarUnverified端到端加密常见实现方案包括基于 X25519 的密钥交换配合 AES-256-GCM 对称加密，局域网直连场景 TLS 1.3 已足够满足传输安全需求73% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1069194API
curl https://kongchang.com/api/v1/knowledge/claims/1069194MCP
get_claim(id=1069194)