Heidy Khlaaf, Chief Scientist at the AI Now Institute and former safety engineer at OpenAI, stated that scientific claims must allow for verification of their truth or falsity. She made this comment regarding discussions on whether AI could lead to the death of all people.
According to Khlaaf, the requirement for verifiability is necessary so that scientific arguments do not take on the character of religious beliefs. The available description of the material contains no additional details about specific systems, scenarios, or conducted research.
The significance of this statement lies in the call to separate verifiable risk assessments from claims that cannot be confirmed or refuted. However, the presented package contains only a synopsis of the AI Now Institute publication and does not provide independent confirmation of the position or its empirical verification.
editorial commentary
Why it matters
The likely consequence is increased demands for the formulation and proof of claims regarding extreme AI risks. The next observable signal would be published verification methods, research results, or independent evaluations of specific scenarios. Significant uncertainty remains: only a brief synopsis is available from the source without such details.