How likely is loss of control over AI? скачать в хорошем качестве

How likely is loss of control over AI?

Annika Hallensleben and Alejandro Ortega (AI Policy Researchers at Apollo Research) discuss chapter 3 of their recent paper about loss of control over AI which looks at ways in which loss of control might materialise in the future. Apollo Research recently published a report on loss of control over AI: The Loss of Control Playbook: Degrees, Dynamics, and Preparedness which aims to make LoC conceptually tractable and operationally useful, so that governments and organizations can start adequately preparing for national security and societal threats from advanced AI systems today. The report does so in 3 steps: proposing new taxonomy to assess loss of control: Deviation, Bounded LoC and Strict LoC Introducing a practical governance framework (“DAP” framework) for straightforward interventions using the levers of deployment context, affordances and permissions that can be actioned today; and, Analysing the long term dynamics that could lead society into a “state of vulnerability” to LoC and the resulting need for adequate preparedness Read the blog post: https://www.apolloresearch.ai/researc... Read the paper: https://arxiv.org/pdf/2511.15846v5