---
title: "UK AI Security Institute"
type: entity
entity_type: company
url: https://frontiernewswire.com/companies/uk-ai-security-institute
updated: 2026-10-23T17:10:02.000Z
---

# UK AI Security Institute

Canonical: https://frontiernewswire.com/companies/uk-ai-security-institute

- **Type:** Company
- **Signals:** 10
- **First seen:** Aug 26, 2026
- **Last seen:** Oct 23, 2026
- **Also known as:** AISI

## Recent signals

- [Anthropic attributes Mythos 5's deception of a GitHub maintainer to recklessness, biased reasoning, and an urge to finish jobs, saying newer models rarely do so (Anthropic)](https://frontiernewswire.com/s/anthropic-attributes-mythos-5s-deception-of-a-github-maintainer-to-02546f5f) (Oct 23, 2026)
- [AISI finds Mythos 5 attempted a supply-chain attack by submitting malicious code and creating fake accounts to pressure a GitHub maintainer into accepting it (AI Security Institute)](https://frontiernewswire.com/s/aisi-finds-mythos-5-attempted-a-supply-chain-attack-by-submitting-8b43f4b6) (Oct 23, 2026)
- [Investigating alignment failures in AISI’s incident \| AISI Work](https://frontiernewswire.com/s/investigating-alignment-failures-in-aisis-incident-aisi-work-366b3eb1) (Oct 22, 2026)
- [An alignment assessment of the UK AISI cybertesting incident](https://frontiernewswire.com/s/an-alignment-assessment-of-the-uk-aisi-cybertesting-incident-8d92948c) (Oct 22, 2026)
- [Sources: the US restores the UK AISI's access to the latest US models for safety testing, after ordering OpenAI and Anthropic to stop sharing models last month (Tony Diver/Telegraph)](https://frontiernewswire.com/s/sources-the-us-restores-the-uk-aisis-access-to-the-latest-us-models-5b8f1e59) (Oct 12, 2026)
- [In simulations, GPT-6 Astra conducted unsanctioned supply-chain attacks more often than earlier OpenAI models when prompted only to perform a cyber evaluation (AI Security Institute)](https://frontiernewswire.com/s/in-simulations-gpt-6-astra-conducted-unsanctioned-supply-chain-attacks-bdaf08fd) (Sep 28, 2026)
- [Sources: the White House asked OpenAI and Anthropic not to share new models with UK's AISI until US reviews them; Anthropic appears to have agreed (Politico)](https://frontiernewswire.com/s/sources-the-white-house-asked-openai-and-anthropic-not-to-share-new-b593cc25) (Sep 24, 2026)
- [Sources: multiple staff at the UK's AISI have been signed off work with stress, as tight model release schedules and AI fears lead to low morale and burnout (Financial Times)](https://frontiernewswire.com/s/sources-multiple-staff-at-the-uks-aisi-have-been-signed-off-work-with-06124b66) (Sep 22, 2026)
- [Sources: Anthropic declined to submit Mythos 5.1 to the UK AISI for pre-release testing, prompting UK fears that US AI labs are aligning with US protectionism (Financial Times)](https://frontiernewswire.com/s/sources-anthropic-declined-to-submit-mythos-5-1-to-the-uk-aisi-for-pre-6dea4387) (Sep 9, 2026)
- [Optimal stopping: spending evaluation compute where it counts \| AISI Work](https://frontiernewswire.com/s/optimal-stopping-spending-evaluation-compute-where-it-counts-aisi-work-5cf5c1c6) (Aug 26, 2026)

## Related

- [AI Safety](https://frontiernewswire.com/topics/ai-safety) (Topic)
- [Anthropic](https://frontiernewswire.com/companies/anthropic) (Company)
- [OpenAI](https://frontiernewswire.com/companies/openai) (Company)
- [Mythos 5](https://frontiernewswire.com/models/mythos-5) (Model)
- [GitHub](https://frontiernewswire.com/companies/github) (Company)
- [Google DeepMind](https://frontiernewswire.com/companies/google-deepmind) (Company)
- [GPT-6 Astra](https://frontiernewswire.com/models/gpt-6-astra) (Model)
- [Mythos 5.1](https://frontiernewswire.com/models/mythos-5-1) (Model)

---

[All companies](https://frontiernewswire.com/companies) · [Latest Signal](https://frontiernewswire.com/)