Home
Technology
Gaming
Entertainment
World & Business
Science
Sports
AI

World & Business

Home
Technology
Gaming
Entertainment
World & Business
Science
Sports
AI

VEPA aims to combine adaptable and deterministic guardrails for AI agents · Digg

Announcement

VEPA aims to combine adaptable and deterministic guardrails for AI agents

Its creators say deployment feedback updates executable safety rules that enforce policies before agents act.

TLDR

VEPA's creators say the guardrail turns safety and security policies into executable symbolic rules that evolve with deployment feedback. They argue LLM-based guardrails can adapt but may make inconsistent decisions, while formal guardrails enforce policies deterministically but are usually fixed. VEPA is intended to check actions before agents execute them, using rules that can be formally verified.

Combined views

716

2 Sources, first seen 2h ago

Combined views

716

2 Sources, first seen 2h ago

26 likes

2h ago

first seen 2h ago

26 likes

5 comments

5 saves

23 reposts

5 comments

5 saves

23 reposts

Featured Source

Sentiment

Positive——Negative

Summary

Not enough discussion yet.

No sentiment analysis available yet.

Sentiment

Positive——Negative

Summary

Not enough discussion yet.

No sentiment analysis available yet.

Today's Rank

—

Not ranked yet

Today's Rank

—

Not ranked yet

2 Sources

@duynguyen772🚨 As agents become more capable, new safety and security risks continue to emerge, as highlighted by recent reports of AI agents taking unauthorized actions or leaking sensitive data. Therefore, we need runtime guardrails that can reliably enforce safety policies before agent actions are executed, while continually adapting to new risks. Existing guardrails usually trade off adaptability and reliability: LLM-based guardrails can adapt with memory, but their decisions may be inconsistent. On the other hand, formal guardrails provide deterministic enforcement but are typically fixed and unable to incorporate new risks that appear after deployment. 🚀 We introduce VEPA, a lifelong formal guardrail for agents. It represents safety and security policies as executable symbolic rules that evolve from deployment feedback. These policies enable deterministic enforcement of agent actions and can be formally verified, avoiding the inconsistency and ambiguity that can occur with natural-language policies. 🧵⬇️2h

@mohitban47RT @duynguyen772: 🚨 As agents become more capable, new safety and security risks continue to emerge, as highlighted by recent reports of AI…2h

2 Sources

@duynguyen772🚨 As agents become more capable, new safety and security risks continue to emerge, as highlighted by recent reports of AI agents taking unauthorized actions or leaking sensitive data. Therefore, we need runtime guardrails that can reliably enforce safety policies before agent actions are executed, while continually adapting to new risks. Existing guardrails usually trade off adaptability and reliability: LLM-based guardrails can adapt with memory, but their decisions may be inconsistent. On the other hand, formal guardrails provide deterministic enforcement but are typically fixed and unable to incorporate new risks that appear after deployment. 🚀 We introduce VEPA, a lifelong formal guardrail for agents. It represents safety and security policies as executable symbolic rules that evolve from deployment feedback. These policies enable deterministic enforcement of agent actions and can be formally verified, avoiding the inconsistency and ambiguity that can occur with natural-language policies. 🧵⬇️2h

@mohitban47RT @duynguyen772: 🚨 As agents become more capable, new safety and security risks continue to emerge, as highlighted by recent reports of AI…2h