What is securitisation

• Securitisation transforms an issue into an "existential threat," justifying extraordinary measures.
• This process is subjective and dependent on framing, e.g., COVID-19 lockdowns, Islamic terrorism surveillance, the Vietnam War.
• National securitisation focuses on state survival, often invoking military actions and power balance.
• Macrosecuritization deals with global existential threats, aiming at protecting humanity as a whole through collaboration.

Sears (2023) - The macrosecuritization of Existential Threats to humanity

• Macrosecuritization failure often results when national securitisation overrides global cooperative measures.
• Successful macrosecuritisations include responses to the ozone hole and nuclear non-proliferation.
• The conflict between national security and humanity security often sees the former prevailing, limiting effective global action.
• National securitisation creates races for supremacy, obstructing global safety strategies like moratoriums or international projects.

Relation to Aschenbrenner's 'Situational Awareness'

• Aschenbrenner's narrative heavily leans on national securitisation, emphasizing U.S. supremacy in AI development.
• He frames China as the existential threat, focusing on the competitive race rather than the AI itself.
• Arguments lack consideration for global strategies, e.g., moratoriums or international collaboration.

Why Aschenbrenner's narrative is dangerous and the role of expert communities

• National securitization narratives, advocated by Aschenbrenner, mimic historical instances that failed to address global threats effectively.
• Aschenbrenner's influence could promote dangerous AI developments prioritizing U.S. dominance over safety.
• Expert communities have historically impacted securitization narratives and can counter Aschenbrenner’s approach.
• The AI Safety community's involvement is crucial in promoting humanity macrosecuritization.

The possibility of a moratorium, military conflict, and collaboration

• Moratoriums on AI development are undervalued by Aschenbrenner, potentially easier without national securitisation.
• National securitisation could lead to military conflicts, including the breakdown of nuclear deterrence.
• Collaboration could be plausible if states perceive AI as destabilizing rather than advantageous, maintaining global peace.
• Effective humanity macrosecuritization can influence enduring international cooperation and safety.