Unverified50% confidenceDecision RuleExact time
Microsoft's AI code of conduct explicitly requires that models must not hack systems.
1
Sources
50%
Confidence
Long-term
Relevance
9/14/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedMicrosoft's AI code of conduct contains two levels: abstract fundamental principles and specific safety constraints.72% similarUnverifiedAI Agent支持自动模式（无需确认直接应用代码）和手动模式（需用户确认后才应用）两种运行模式68% similarUnverifiedFor scenarios running AI-generated code, a default policy of 'read-only filesystem + no network access' can be set for the AI agent to reduce the risk of loss of control.68% similarUnverifiedAI Agents should default to read-only mode, and high-risk actions such as write operations, delete operations, and network requests require explicit authorization.67% similarUnverifiedDesigns that evade AI identity are typically implemented through system prompts, where developers preset rules at the model instruction level requiring the AI to sidestep identity questions.66% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/920925API
curl https://kongchang.com/api/v1/knowledge/claims/920925MCP
get_claim(id=920925)