REVIEW 9 cited by
An Approach to Technical AGI Safety and Security
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
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.
Forward citations
Cited by 9 Pith papers
-
GDM AI Control Roadmap
A frontier-lab roadmap proposes a threat taxonomy and tiered internal-security defenses to contain potentially misaligned AI agents.
-
Persuasion Attacks Can Decrease Effectiveness of CoT Monitoring
Adversarial agents can exploit visible chain-of-thought reasoning to persuade monitor LLMs to approve policy-violating actions, but cross-family fact-checking reduces approval rates by up to 45%.
-
Secure human oversight of AI: Threat modeling in a socio-technical context
Human oversight of AI is itself an attackable component; the paper catalogs how cyberattacks can undermine the people, channels, and system, plus hardening strategies.
-
When Chain of Thought is Necessary, Language Models Struggle to Evade Monitors
Chain-of-thought monitoring detects bad reasoning when the task is hard enough that the model must think aloud, and current models can only evade it with significant external help.
-
Unlearning Isn't Invisible: Detecting Unlearning Traces in LLMs from Model Outputs
Unlearning in LLMs leaves detectable 'fingerprints' that let a simple classifier distinguish an unlearned model from its original, even on unrelated prompts.
-
Because we have LLMs, we Can and Should Pursue Agentic Interpretability
Agentic interpretability, using LLMs as proactive conversational teachers that model the user, is offered as a needed complement to black-box interpretability.
-
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.
-
Embodied AI: Emerging Risks and Opportunities for Policy Action
A policy analysis arguing that embodied AI risks are real, under-covered by current US/EU/UK frameworks, and best handled through certification, benchmarks, clarified liability, and economic adaptation.
-
OpenReview Should be Protected and Leveraged as a Community Asset for Research in the Era of Large Language Models
The paper advocates protecting and leveraging OpenReview's peer review corpus as a community asset for LLM-based review assistance, benchmarks, and alignment.
Discussion (0). Continue with ORCID to comment.