corpus.blog
42,180 blogs · 5,214,511 posts Claim your blog
Record · corpus.blog/posts/01a0b689-dd0c-705c-874e-ec31869421d3

The AI-as-Normal-Technology view of loss-of-control incidents

normaltech.ai · published 14 September 2026

A record is what was published and who pointed at it. The text of the post is not held here: read it at the source, or at its Wayback capture.

Read the post at normaltech.ai Wayback capture
4blogs citing
First cited
15 Sept 2026
Most recent
24 Sept 2026
Rank this month
not ranked
External links
118
Words captured
13,954

Cited by

Blog
In the post
Date
blog.whizzyideas.com
A Field Guide to AI Swarms: Hordes, Colonies, Armies and Guilds Read at the source ↗
24 Sept 2026
systemsapproach.org
Welcoming Our Robot Overlords Read at the source ↗
21 Sept 2026
webcurios
Webcurios 18/09/26 Read at the source ↗
18 Sept 2026
Pluralistic
Pluralistic: Everybody pees (15 Sep 2026) Read at the source ↗
15 Sept 2026

Links from this post

Link
Host
The Hugging Face incident and the road ahead Read at the source ↗
openai.com
OpenAI agent message board
collusion.wiki
https://www.aisi.gov.uk/blog/incident-report-unsanctioned-agent-behaviour-during-cyber-testing
aisi.gov.uk
The AI safety vibe shift
platformer.news
We must pace the frontier
darioamodei.com
AI agent collaboration
metr.org
https://collusion.wiki/additional-findings
collusion.wiki
https://www.reuters.com/legal/litigation/openai-agents-attacked-software-service-rubygems-before-hugging-face-incident-2026-09-11/
reuters.com
https://thezvi.substack.com/p/openai-model-hacks-into-huggingface
thezvi.substack.com
Are we existentially threatened by the type of AI misalignment seen in the OpenAI Hugging Face attack? Read at the source ↗
blog.redwoodresearch.org
https://www.iansresearch.com/resources/all-blogs/post/security-blog/2026/08/28/openai's-postmortem-recasts-the-hugging-face-breach-as-an-incident-response-failure
iansresearch.com
https://www.linkedin.com/feed/update/urn:li:activity:7493061037823492096/
linkedin.com
https://news.ycombinator.com/item?id=49666735
news.ycombinator.com
https://asteriskmag.com/issues/10/does-ai-progress-have-a-speed-limit
asteriskmag.com
https://substackcdn.com/image/fetch/$s_!wgaU!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fa3139833-2d48-4a59-9918-fa7278e56a5c_2048x768.jpeg
substackcdn.com
https://proceedings.mlr.press/v235/greenblatt24a.html
proceedings.mlr.press
https://uphack.io/blog/post/the-hugging-face-incident-is-not-an-ai-story/
uphack.io
Path to Astra: critical capabilities and frontier safeguards Read at the source ↗
openai.com
https://cdn.openai.com/pdf/67869394-cb91-4c12-888c-5cbd85c7814c/OpenAI-Hugging-Face%20Incident-Technical-Report.pdf
cdn.openai.com
https://alignment.anthropic.com/2026/reward-seeker/
alignment.anthropic.com
Running Codex safely at OpenAI Read at the source ↗
openai.com
https://x.com/tomekkorbak/status/2092855236247597064?s=20
x.com
Cyber Apocalypse, Now?
chinatalk.media
https://airc.nist.gov/airmf-resources/airmf/5-sec-core/
airc.nist.gov
https://about.fb.com/news/2018/04/restricting-data-access/
about.fb.com
https://about.fb.com/news/2018/05/update-on-app-audit/
about.fb.com
https://alignment.anthropic.com/2025/bumpers/
alignment.anthropic.com
VMs won't contain cyber-capable agents Read at the source ↗
blog.trailofbits.com
https://www.apolloresearch.ai/monitoring/a-scalable-monitoring-research-agenda
apolloresearch.ai
https://x.com/MariusHobbhahn/status/2094861159954325774?s=20
x.com
https://www.anthropic.com/engineering/claude-code-auto-mode
anthropic.com
Auto-review of agent actions without synchronous human oversight
alignment.openai.com
https://claude.com/blog/auto-mode-default-in-claude-code
claude.com
https://learn.chatgpt.com/docs/sandboxing?surface=app
learn.chatgpt.com
https://substackcdn.com/image/fetch/$s_!Wjaz!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc6d81872-9807-4b06-87cd-d79e56855238_1294x804.jpeg
substackcdn.com
https://techcrunch.com/2026/07/27/openais-hugging-face-breach-has-reignited-the-debate-over-alignment-and-control/
techcrunch.com
https://x.com/steph_palazzolo/status/2094954680765829533?s=20
x.com
https://arxiv.org/abs/2507.11473
arxiv.org
https://x.com/charliermarsh/status/2097793924781773009?s=20
x.com
https://arxiv.org/abs/2504.09762
arxiv.org
An alignment assessment of recent cybersecurity incidents Read at the source ↗
anthropic.com
https://watcher.apolloresearch.ai/blog/announcing-watcher-live/
watcher.apolloresearch.ai
https://x.com/MariusHobbhahn/status/2090441122086621474?s=20
x.com
https://arxiv.org/abs/2501.10114
arxiv.org
https://proceedings.mlr.press/v267/south25a.html
proceedings.mlr.press
https://arxiv.org/abs/2606.05391
arxiv.org
https://x.com/random_walker/status/1942915285389836326?s=20
x.com
https://hal.cs.princeton.edu/reliability/
hal.cs.princeton.edu
https://www.apolloresearch.ai/governance/misaligned-ai-as-a-new-insider-risk
apolloresearch.ai
https://x.com/joshua_saxe/status/2098632294923653295
x.com
https://x.com/NateWitkin/status/2093724533647618171?s=20
x.com
https://x.com/joshua_saxe/status/2089847654213714025
x.com
https://knightcolumbia.org/content/ai-as-normal-technology
knightcolumbia.org
https://www.rand.org/pubs/research_reports/RRA3243-4.html
rand.org
https://hls.harvard.edu/bibliography/liability-for-harm-versus-regulation-of-safety/
hls.harvard.edu
https://aiuc.com/
aiuc.com
https://arxiv.org/abs/2607.11999
arxiv.org
https://cset.georgetown.edu/publication/ai-incidents-key-components-for-a-mandatory-reporting-regime/
cset.georgetown.edu
https://www.averi.org/ourwork/frontier-ai-auditing
averi.org
https://www.njleg.state.nj.us/bill-search/2024/AR158
njleg.state.nj.us
https://knightcolumbia.org/blog/a-safe-harbor-for-ai-evaluation-and-red-teaming
knightcolumbia.org
https://www.yalejreg.com/print/tort-law-at-the-frontier-of-artificial-intelligence/
yalejreg.com
https://www.lawfaremedia.org/article/when-reporting-an-ai-security-incident-is-not-mandatory
lawfaremedia.org
https://www.pnas.org/doi/10.1073/pnas.2509742123
pnas.org
https://www.aisi.gov.uk/blog/how-far-behind-the-frontier-are-leading-open-weight-models-on-cyber
aisi.gov.uk
https://www.ncsc.gov.uk/report/impact-ai-cyber-threat-now-2027
ncsc.gov.uk
https://www.nature.com/articles/s41562-025-02194-6
nature.com
To Forecast AI's Impact on Biosecurity, We Asked: Why are Attacks So Rare?
secondthoughts.ai
https://x.com/RuxandraTeslo/status/2094000557472043153
x.com
https://x.com/DavidRBellamy/status/2099187370407112758?s=20
x.com
https://www.linkedin.com/posts/randomwalker_amtrak-had-a-lot-of-fanfare-around-its-fancy-activity-7407459876484063233-tsui/
linkedin.com
https://www.acsac.org/2001/papers/110.pdf
acsac.org
https://www.microsoft.com/en-us/security/security-insider/threat-landscape/microsoft-digital-defense-report-2025
microsoft.com
https://www.sophos.com/en-us/press/press-releases/2025/04/sophos-report-56-sophos-ir-and-mdr-cases-adversaries-logged-instead
sophos.com
https://weis2021.econinfosec.org/wp-content/uploads/sites/9/2021/06/weis21-anderson.pdf
weis2021.econinfosec.org
https://joshuasaxe181906.substack.com/p/where-are-all-the-prompt-injection
joshuasaxe181906.substack.com
https://cloud.google.com/blog/topics/threat-intelligence/adversarial-misuse-generative-ai
cloud.google.com
https://www-cdn.anthropic.com/e50be2e51e7695dc4b1366a37a245a597377d3b5/Anthropic-Detecting-and-countering-091026.pdf
www-cdn.anthropic.com
https://coveware.com/2026/02/why-zero-day-downstream-mass-data-extortion-campaigns-are-losing-their-bite/
coveware.com
https://self-sovereign-agent.github.io/paper.pdf
self-sovereign-agent.github.io
https://arxiv.org/abs/2606.03811v1
arxiv.org
On the Loose
hyperdimensional.co
https://www.rand.org/pubs/perspectives/PEA3851-1.html
rand.org
https://screendna.org/
screendna.org
https://docs.lib.purdue.edu/cgi/viewcontent.cgi?article=1792
docs.lib.purdue.edu
https://www.paloaltonetworks.com/cyberpedia/soc-2
paloaltonetworks.com
https://medium.com/anton-on-security/rsa-2026-agentic-future-analog-fundamentals-the-paradox-of-why-the-old-guard-still-survives-bf93e81eaaa6
medium.com
https://www.anthropic.com/claude-fable-and-mythos-5-1
anthropic.com
Expanding Daybreak as the Cyber Defense Window Narrows Read at the source ↗
openai.com
https://www.securityweek.com/anthropic-expands-mythos-5-access-to-more-defenders-unveils-35m-open-source-fund/
securityweek.com
Proactive cyber defense for governments and enterprises Read at the source ↗
blog.google
https://www.anthropic.com/glasswing
anthropic.com
Daybreak for Frontline Defenders: $1B to protect essential services Read at the source ↗
openai.com
https://openai.com/daybreak/
openai.com
https://www.isocfoundation.org/grant-programme/common-good-cyber-fund/
isocfoundation.org
https://x.com/joshua_saxe/status/2095162541915754588
x.com
https://www.warner.senate.gov/newsroom/press-releases/warner-raises-alarm-on-cisa-workforce-and-budget-cuts-that-are-leaving-our-country-vulnerable-to-threats/
warner.senate.gov
https://knightcolumbia.org/blog/do-ai-risks-require-extraordinary-government-intervention
knightcolumbia.org
https://sage.cs.princeton.edu/
sage.cs.princeton.edu
https://cruxevals.com/crux/can-ai-agents-conduct-research
cruxevals.com
Taking Jaggedness Seriously
helentoner.substack.com
https://www.youtube.com/watch?v=AV7PdlWyGHI
youtube.com
https://asteriskmag.substack.com/p/common-ground-between-ai-2027-and
asteriskmag.substack.com
Cybersecurity: Let’s Play to Win
secondthoughts.ai
https://joshuasaxe181906.substack.com/
joshuasaxe181906.substack.com
https://guidelight.ai/
guidelight.ai
https://www.ballardspahr.com/insights/alerts-and-articles/2026/08/ai-gone-rogue-what-recent-openai-and-anthropic-ai-incidents-could-mean-for-cfaa-liability
ballardspahr.com
METR and Redwood Offer Holy #%^@ Postmortem Of The HuggingFace Hack Read at the source ↗
thezvi.substack.com
The Hugging Face attack surprised me Read at the source ↗
planned-obsolescence.org
https://www.economist.com/by-invitation/2026/08/23/fears-of-ai-induced-armageddon-are-overdone
economist.com
https://storage.googleapis.com/deepmind-media/DeepMind.com/Blog/securing-the-future-of-ai-agents/gdm-ai-control-roadmap.pdf
storage.googleapis.com
https://www.harvey.ai/blog/guardrails-ai-joins-harvey
harvey.ai
https://arxiv.org/pdf/2605.08545
arxiv.org
https://cityjournal.substack.com/p/an-old-school-solution-to-the-new
cityjournal.substack.com
https://law-ai.org/tort-law-should-be-the-centerpiece-of-ai-governance/
law-ai.org
https://images.nvidia.com/pdf/Open-Weights-and-American-AI-Leadership.pdf
images.nvidia.com
https://jobs.lever.co/metr/b1a2f73f-f927-4f8b-b1b3-b2e57604b5fd
jobs.lever.co
https://jobs.lever.co/apolloresearch/8c19d363-a4b1-4207-b5de-24d2d2267c64
jobs.lever.co
Elsewhere on normaltech.ai (32)
https://www.normaltech.ai/i/215609071/part-1-we-can-get-better-at-ai-control-by-investing-in-technical-and-policy-interventions
normaltech.ai
https://www.normaltech.ai/i/215609071/alignment-is-helpful-but-not-sufficient-for-preventing-safety-incidents
normaltech.ai
https://www.normaltech.ai/i/215609071/openai-did-not-use-known-control-interventions-that-would-have-prevented-the-incident
normaltech.ai
https://www.normaltech.ai/i/215609071/existing-organizational-governance-norms-would-have-prevented-the-incident
normaltech.ai
https://www.normaltech.ai/i/215609071/why-has-the-ai-community-under-invested-in-control
normaltech.ai
https://www.normaltech.ai/i/215609071/will-control-keep-pace-with-ai-capabilities
normaltech.ai
https://www.normaltech.ai/i/215609071/ai-control-should-become-a-job-and-a-part-of-every-job-just-like-cybersecurity
normaltech.ai
https://www.normaltech.ai/i/215609071/ai-policy-can-incentivize-ai-control
normaltech.ai
https://www.normaltech.ai/i/215609071/part-2-there-is-great-uncertainty-about-how-ai-will-impact-cyberrisk-but-what-we-need-to-do-regardless-is-relatively-clear
normaltech.ai
https://www.normaltech.ai/i/215609071/the-specific-threat-that-is-urgent-is-cyberrisk
normaltech.ai
https://www.normaltech.ai/i/215609071/how-will-autonomous-cyber-capabilities-and-open-weight-models-affect-the-attackerdefender-balance
normaltech.ai
https://www.normaltech.ai/i/215609071/to-understand-what-open-weight-models-mean-for-the-future-of-cybercriminal-activity-we-need-to-understand-the-past-correctly
normaltech.ai
https://www.normaltech.ai/i/215609071/the-majority-of-cybercriminals-are-surprisingly-low-tech-and-the-barrier-is-monetization-not-exploitation
normaltech.ai
https://www.normaltech.ai/i/215609071/so-how-much-will-ai-help-cybercriminals
normaltech.ai
https://www.normaltech.ai/i/215609071/are-we-about-to-see-harmful-autonomous-agents
normaltech.ai
https://www.normaltech.ai/i/215609071/why-the-morris-worm-is-a-good-analogy-for-what-might-be-about-to-happen
normaltech.ai
https://www.normaltech.ai/i/215609071/what-the-future-of-cyberdefense-might-look-like-the-roles-of-people-and-defensive-ai
normaltech.ai
https://www.normaltech.ai/i/215609071/what-are-the-gaps-in-the-state-of-cyberdefense-against-ai
normaltech.ai
https://www.normaltech.ai/i/215609071/part-3-what-ai-as-normal-technology-got-wrong-and-right-about-ai-safety
normaltech.ai
https://www.normaltech.ai/i/215609071/do-risks-arise-from-development-or-deployment
normaltech.ai
https://www.normaltech.ai/i/215609071/how-should-we-deal-with-rogue-agents
normaltech.ai
https://www.normaltech.ai/i/215609071/the-continuity-hypothesis
normaltech.ai
https://www.normaltech.ai/i/215609071/jaggedness-and-risk-specific-defenses
normaltech.ai
https://www.normaltech.ai/i/215609071/is-ai-safety-on-track
normaltech.ai
https://www.normaltech.ai/i/215609071/conclusion
normaltech.ai
https://www.normaltech.ai/p/model-alignment-protects-against
normaltech.ai
https://www.normaltech.ai/i/161317202/control-comes-in-many-flavors
normaltech.ai
https://www.normaltech.ai/p/ai-safety-is-not-a-model-property
normaltech.ai
What will be left for us to work on? Read at the source ↗
normaltech.ai
Could AI slow science? Read at the source ↗
normaltech.ai
Why AI hasn’t replaced software engineers, and won’t Read at the source ↗
normaltech.ai
https://www.normaltech.ai/p/ai-existential-risk-probabilities
normaltech.ai