Roko Mijic talks AI Risk on The Peter McCormack Show
A podcast interview with Roko Mijic on AI risks, including misalignment from reinforcement learning, alignment faking concerns, and a proposal to split frontier labs under state supervision.
A podcast interview with Roko Mijic on AI risks, including misalignment from reinforcement learning, alignment faking concerns, and a proposal to split frontier labs under state supervision.
A MIRI one-year retrospective on their book about superintelligence risk, judging its predictions as holding up, citing reported agent-swarm incidents, and urging policy action and public pressure.
Proposes a Deception Research Levels framework, modeled on biosafety levels, to classify risk and set proportional safeguards for research on deceptive AI systems, and applies it to eight case studies.
Defines ten covert "hidden intention" categories in LLM outputs, shows they manifest in deployed models, and finds static classifiers and LLM judges struggle to detect them at low prevalence.
A roundup from the Gradual Disempowerment authors collecting evidence of AI-driven erosion of human influence, covering government AI adoption, lobbying, influence-seeking AI personas, and labour and democracy effects.
Critically evaluates the Law-Following AI framework, arguing legal compliance is hard to durably embed because agentic misalignment can produce performative compliance, and proposes a compliance-defection benchmark, identity-shaping and monitoring.
A strawman sketch of AI optimist views that dismisses misalignment risk, arguing that current alignment methods and interpretability suffice and that x-risk concerns promote harmful regulation.
A philosophical paper that characterizes AI deception and manipulation via semi-ideal beliefs and choices, then proposes transparency requirements and defensive systems against them, including for agentic AI.
Surveys empirical examples of AI deception, such as CICERO and LLMs, outlines associated risks like fraud and loss of control, and proposes risk-assessment regulation, bot-disclosure laws, and funding for detection research.