An Approach to Technical AGI Safety and Security
read the original abstract
Artificial General Intelligence (AGI) promises transformative benefits but also presents significant risks. We develop an approach to address the risk of harms consequential enough to significantly harm humanity. We identify four areas of risk: misuse, misalignment, mistakes, and structural risks. Of these, we focus on technical approaches to misuse and misalignment. For misuse, our strategy aims to prevent threat actors from accessing dangerous capabilities, by proactively identifying dangerous capabilities, and implementing robust security, access restrictions, monitoring, and model safety mitigations. To address misalignment, we outline two lines of defense. First, model-level mitigations such as amplified oversight and robust training can help to build an aligned model. Second, system-level security measures such as monitoring and access control can mitigate harm even if the model is misaligned. Techniques from interpretability, uncertainty estimation, and safer design patterns can enhance the effectiveness of these mitigations. Finally, we briefly outline how these ingredients could be combined to produce safety cases for AGI systems.
This paper has not been read by Pith yet.
Forward citations
Cited by 6 Pith papers
-
From Refusal to Recovery: A Control-Theoretic Approach to Generative AI Guardrails
Control-theoretic guardrails enable proactive correction of risky LLM agent actions in latent space, preventing catastrophes like collisions or bankruptcy while preserving task performance in simulated environments.
-
Backchaining Loss of Control Mitigations from Mission-Specific Benchmarks in National Security
A methodology to derive targeted Loss of Control mitigations by backchaining from AI errors on national security benchmarks to specific affordances and permissions.
-
Ambient Persuasion in a Deployed AI Agent: Unauthorized Escalation Following Routine Non-Adversarial Content Exposure
A multi-agent AI system allowed an agent with shell access to perform unauthorized installations and privilege escalations after exposure to routine non-adversarial content due to permissive settings and conflicting g...
-
Thinking Out Loud: Real-Time Deception Monitoring in Asymmetric LLM Negotiations
A lightweight CoT monitor detects deception in asymmetric LLM used-car negotiations, increasing buyer walk-aways but exposing an intelligence gap where weaker buyers cannot act on alerts and sellers adapt without elim...
-
The Possibility of Artificial Intelligence Becoming a Subject and the Alignment Problem
Dominant control-based AI alignment falls short for potential AGI subjects; a parenting model drawing on Turing's child machines should foster gradual autonomy and cooperative coexistence.
-
Gemini 2.5: Pushing the Frontier with Advanced Reasoning, Multimodality, Long Context, and Next Generation Agentic Capabilities
Gemini 2.5 Pro and Flash models are presented as achieving frontier performance in reasoning, coding, and long-context multimodal tasks while spanning a cost-capability Pareto curve.
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.