Com­puter Se­cu­rity & Cryptography

End-to-end hack­ing with lan­guage models

11 dice­ware words is enough

Prevent­ing model exfil­tra­tion with up­load limits

How im­por­tant is AI hack­ing as LLMs ad­vance?

In­cor­po­rat­ing Mechanism De­sign Into De­ci­sion Theory

Pro­tect­ing agent boundaries

Agent mem­branes and causal distance

Plan­ning to build a cryp­to­graphic box with perfect secrecy

Nav­i­gat­ing the Attackspace

n of m ring signatures

Ap­ply to the Con­cep­tual Boundaries Work­shop for AI Safety

[Question] How much should e-sig­na­tures have to cost a coun­try?

5 Rea­sons Why Govern­ments/​Mili­taries Already Want AI for In­for­ma­tion Warfare

AI Safety is Drop­ping the Ball on Clown Attacks

Fix­ing In­sider Threats in the AI Sup­ply Chain

Re­vis­it­ing the Man­i­fold Hypothesis

Biose­cu­rity Cul­ture, Com­puter Se­cu­rity Culture

The rise of AI in cybercrime

Pri­vacy in a Digi­tal World

Fron­tier Model Security

Proof of pos­te­ri­or­ity: a defense against AI-gen­er­ated misinformation

gamers be­ware: mod­ded Minecraft has new malware

Trans­former Ar­chi­tec­ture Choice for Re­sist­ing Prompt In­jec­tion and Jail-Break­ing Attacks

Microsoft and Google us­ing LLMs for Cybersecurity

[Question] AI in­ter­pretabil­ity could be harm­ful?

Prevent­ing AI Mi­suse: State of the Art Re­search and its Flaws

Up­skil­ling, bridge-build­ing, re­search on se­cu­rity/​cryp­tog­ra­phy and AI safety

Cryp­to­graphic and aux­iliary ap­proaches rele­vant for AI safety

The Virus—Short Story

Boundaries-based se­cu­rity and AI safety approaches

Reli­a­bil­ity, Se­cu­rity, and AI risk: Notes from in­fosec text­book chap­ter 1

AI in­fosec: first strikes, zero-day mar­kets, hard­ware sup­ply chains, adop­tion barriers

POC || GTFO cul­ture as par­tial an­ti­dote to al­ign­ment wordcelism

[Question] Hoard­ing Gmail-ac­counts in a post-CAPTCHA world?

AI Gover­nance & Strat­egy: Pri­ori­ties, tal­ent gaps, & opportunities

The 2/​3 rule for multi-fac­tor authentication

How Likely is Los­ing a Google Ac­count?

AI se­cu­rity might be helpful for AI alignment

Pri­vacy Tradeoffs

Not Get­ting Hacked

Hacker-AI and Cy­ber­war 2.0+

Non-Tech­ni­cal Prepa­ra­tion for Hacker-AI and Cy­ber­war 2.0+

Us­ing GPT-Eliezer against ChatGPT Jailbreaking

Safe Devel­op­ment of Hacker-AI Coun­ter­mea­sures – What if we are too late?

Hacker-AI – Does it already ex­ist?

Why do we post our AI safety plans on the In­ter­net?

[Question] What is good Cy­ber Se­cu­rity Ad­vice?

Deep­fake(?) Phishing

Im­proved Se­cu­rity to Prevent Hacker-AI and Digi­tal Ghosts

Hacker-AI and Digi­tal Ghosts – Pre-AGI

Se­cu­rity Mind­set: Les­sons from 20+ years of Soft­ware Se­cu­rity Failures Rele­vant to AGI Alignment

Crypto-fed Computation

[Question] How much does cy­ber­se­cu­rity re­duce AI risk?

The Un­bear­able Light­ness of Web Vulnerabilities

[Question] Is there a con­ve­nient way to make “sealed” pre­dic­tions?

[Question] Trans­fer­ring cre­dence with­out trans­fer­ring ev­i­dence?

En­tropy isn’t suffi­cient to mea­sure pass­word strength

Lat­a­cora might be of in­ter­est to some AI Safety organizations

Se­cure homes for digi­tal people

[Question] How good is se­cu­rity for LessWrong and the Align­ment Fo­rum?

Bet­ter Pass­word Peppering

Paper re­view: A Cryp­to­graphic Solu­tion to a Game The­o­retic Problem

The dan­ger of er­ror descriptions

[Question] In soft­ware en­g­ineer­ing, what are the up­per limits of Lan­guage-Based Se­cu­rity?

On De­stroy­ing the World

On See­ing Through ‘On See­ing Through: A Unified The­ory’: A Unified Theory

Scry­ing for out­comes where the prob­lem of deep­fakes has been solved

Uniker­nels: No Longer an Aca­demic Exercise

Es­ti­mat­ing the con­se­quences of de­vice de­tec­tion tech

Neu­ral pro­gram syn­the­sis is a dan­ger­ous technology

Se­cu­rity Mind­set and the Lo­gis­tic Suc­cess Curve

Se­cu­rity Mind­set and Or­di­nary Paranoia

Work on Se­cu­rity In­stead of Friendli­ness?

Op­ti­mal User-End In­ter­net Se­cu­rity (Or, Ra­tional In­ter­net Brows­ing)

Ho­mo­mor­phic en­cryp­tion and Bitcoin

Crypt­anal­y­sis as Episte­mol­ogy? (pag­ing cryp­ton­erds)

Com­puter se­cu­rity story

What is Cryp­to­graph­i­cally Possible

Cryp­to­graphic Boxes for Un­friendly AI