source · primary doc

FLI AI Safety Index 2024

src_fli_ai_safety_index_2024

https://futureoflife.org/document/fli-ai-safety-index-2024/

reliability

0.85

authors: Future of Life Institute

published: 2024-12-11

accessed: 2026-04-19

Notes

FLI-published index with named independent expert panel; primary document of the evaluation itself. Slightly below primary_doc prior because FLI is an advocacy org with a stated position on AI risk.

Intake provenance

method: httpx
tool: afls-ingest/0.0.1
git sha: 4d098737f648
at: 2026-04-19T23:07:10.299958Z
sha256: a56b7764b9bc…

Evidence from this source (5)

Per panelist Stuart Russell, none of the safety activity at the six evaluated c… support

weight

0.85

method: expert_estimate · locator: Panellist Comments (Stuart Russell)

“none of the current activity provides any kind of quantitative guarantee of safety; nor does it seem possible to provide such guarantees given the current approach to AI via giant black boxes”
The FLI AI Safety Index 2024 panel judged the current strategies of all six eva… support

weight

0.92

method: expert_estimate · locator: Key Findings: Control-Problem

“the review panel deemed the current strategies of all companies inadequate for ensuring that these systems remain safe and under human control”
The FLI AI Safety Index 2024 panel found that all flagship models from the six… support

weight

0.90

method: expert_estimate · locator: Key Findings: Jailbreaks

“All the flagship models were found to be vulnerable to adversarial attacks.”
The 2024 FLI AI Safety Index convened an independent panel of seven AI and gove… support

weight

0.95

method: primary_testimony · locator: Methodology section

“An independent review panel of leading experts... volunteered to assess the companies' performances across 42 indicators of responsible conduct”
The FLI AI Safety Index 2024 review panel consists of Yoshua Bengio (Université… support

weight

0.98

method: primary_testimony · locator: Independent Review Panel section