Safety Features for a Centralized AGI Project

Sarah Hastings-Woodhouse · 2025

Read Safety Features for a Centralized AGI Project by Sarah Hastings-Woodhouse (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2507-21082.txt.

Description

Sarah Hastings-Woodhouse examines what safety steps, in a 2025 report, a government-run project to build very advanced AI would need. She starts from the claim that AI may soon match or beat human skill across most mental tasks. Some experts place that point within this decade. Such a system could pose grave threats to national and global security. Today, most AI work sits in private firms with little outside oversight.

The report studies a scenario where the United States government takes direct control of development. It names four high-level priorities for that project. It then sets out seven safety features meant to cut the risks. The features respond to the dangers of racing ahead, losing control, and misuse. The paper offers the plan as a guide for policy debate, not as a claim that central control is certain or simple.

Catalog record

CREATORSarah Hastings-Woodhouse
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE84 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2507-21082

Browse the full library · About the Tower of Babel