Enabling External Scrutiny of AI Systems

2025

Read Enabling External Scrutiny of AI Systems (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-2502-05219.txt.

Description

Kendrea Beers and Helen Toner describe how privacy enhancing tools can open artificial intelligence systems to outside scrutiny, in a 2025 article. Outside checks give the public a window into how AI is built, yet firms resist access for sound reasons: security, privacy, and trade secrets. The nonprofit OpenMined has built shared infrastructure to bridge that gap.

Two real cases anchor the article. One probed social media recommendation systems under a Christchurch Call project, and the other let a British safety institute test frontier models. The authors map which kinds of scrutiny current setups allow and what future setups could add. They conclude the approach merits wider support. Policy makers, they suggest, can back researchers on the legal side as well.

Catalog record

YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE14 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2502-05219

Browse the full library · About the Tower of Babel