Claim · 2023-06-24 · publication · Stated in source · Supported

Sandbrink’s article (arXiv v1, 24 June 2023) lists, among measures against biological misuse of large language models, releasing powerful models only with structured access and considering limits on open sourcing.

One recorded claim. The claim panel gives its qualification, the source passages with their place in each source, and how it was checked.

What it connects

Artificial intelligence and biological misuse: Differentiating risks of language models and biological design tools → articulates → Public release of model weights increases biological misuse risk

Where it appears

Arguments about AI and biological risk →

Claim and evidence

Artificial intelligence and biological misuse: Differentiating risks of language models and biological design toolsarticulates ↓Public release of model weights increases biological misuse risk

Machine-checkedStated in sourceSupported

Supporting evidence: primary · 1 source

Subject version: arXiv v1 (24 June 2023) · Published 2023-06-24 · Effective date not established · Revised 2023-12-23
Date: 2023-06-24 · publication Records dated 2023 →
Role: Earliest located
Register (kind of text): Nonfiction

Scope and limitsA recommendation in a table of mitigations; it does not quantify the risk of open release. Earliest located means the earliest text found by the collection’s searches and the cross-family re-check of 2026-11-02, not a first statement.

The evidence 01

primary sourceSupports the claim

Artificial intelligence and biological misuse: Differentiating risks of language models and biological design tools

arXiv (preprint, Jonas B. Sandbrink, Nuffield Department of Medicine, University of Oxford) · Publication date unknown

Cite the original source

Cite the original document, not Selekoa. Check these details against it.

Source excerpt

“Only release powerful LLMs with structured access; consider limits on open sourcing and strengthen information security to prevent leaks”

Location in source Table 1, row “Risk mitigation”

How it was read Retrieved directly from the publisher

Machine-checked · 2026-11-02

Source details and check notes

Authors: arXiv (preprint, Jonas B. Sandbrink, Nuffield Department of Medicine, University of Oxford)
Version: Retrieved for the candidate pack on 2026-10-28 (SHA-256 1ab85d8647f2d23079d35eb8162b70b3f970094106808ef4291fb89615db0afa). arXiv v1 of 24 June 2023 (the article was revised to v8 on 23 December 2023; this record cites v1 only) Retrieved again at admission on 2026-10-30 (SHA-256 1ab85d8647f2d23079d35eb8162b70b3f970094106808ef4291fb89615db0afa, the same bytes).
Canonical address: Open the canonical source ↗
Accessed: 2026-10-28
Language: English (en)
Attribution: arXiv (preprint, Jonas B. Sandbrink, Nuffield Department of Medicine, University of Oxford), “Artificial intelligence and biological misuse: Differentiating risks of language models and biological design tools”.
Reuse: No open redistribution licence asserted. Bibliographic metadata, editorial paraphrases and short attributed quotations (at most 50 words each). Recheck terms before wider redistribution.
Check notes: Quotation matched on 2026-11-02 against the held source, retrieved on 2026-11-02 from the source’s evidence URL (SHA-256 1ab85d8647f2d23079d35eb8162b70b3f970094106808ef4291fb89615db0afa). Matching ignored differences of whitespace, case, quotation marks, dashes and Unicode compatibility forms; “…” joins fragments that occur in order. Cross-family pass: the re-check of 2026-11-02 (review/earliest-located-recheck-2026-11-02.md, role 170) read this text for this proposition and role. Admitted by the maintainer’s decision of 2026-11-02. No human editorial review.

Status and meaning

Editorial status. Machine-checked. Machines found this record and checked it against its sources; no person accepted it individually. Each evidence record above shows how it was checked.

Relation. A passage of the document meets the proposition’s criterion. The record documents what the text states, in its register (fiction and satire are labelled); it is not evidence that the proposition is true and implies no influence on or from another text.

Date. Date of the arXiv v1 submission; the article was revised to v8 on 23 December 2023, and only v1 is cited.

Record ID ab-arg-d6-sandbrink-articulates