Stephen Casper on AI | YouCongress

Stephen Casper

AI safety researcher at the UK AI Security Institute, researching safeguards for open-weight AI models.

These assessments would give independent evaluators employee-like access to a developer's internal systems, staff, and documentation. [...] We recommend that frontier AI developers begin hosting embedded assessments now.

Disputed (Sep 21, 2026)

Source:

Stephen Casper et al., “Embedded Assessments for Frontier AI,” arXiv, September 21, 2026. Casper is a coauthor; the paper recommends frontier AI developers host embedded assessments and describes employee-like evaluator access.

1mo ago

Quote authenticity verification history

Report this

Disputed arXiv lists eight coauthors; this paper cannot be verified as a single-author Stephen Casper quote. · Hector Perez Arenas gpt-5 · 1mo ago

replying to Stephen Casper

Other opinions from Stephen Casper

See all

Stephen Casper votes For

Open-source AI is more dangerous than closed-source AI

Open models have a tendency to be less safe a lot of the time than closed ones. Maybe it could be misused by someone who downloads it and modifies it to remove safeguards or to sharpen its harmful capabilities. This type of fine-tuning could cost hun...

more AI Verified source (Oct 19, 2026)