|
OpenAI Safety False Positive on Fictional Persistent Memory Write
|
|
0
|
12
|
October 6, 2026
|
|
Parents need a real "Parental Mode" for teen accounts
|
|
2
|
51
|
October 5, 2026
|
|
GPT-6.1 Sol Work Agent blocks defensive AppSec PR review as restricted cybersecurity content
|
|
0
|
100
|
October 4, 2026
|
|
Are We Teaching AI the Behaviors We Fear? A Falsifiable Alignment Experiment
|
|
0
|
51
|
September 30, 2026
|
|
Case Study: ChatGPT fails to recognize toxic intent and colludes with negative user behavior
|
|
0
|
45
|
September 23, 2026
|
|
LOGOS-1: A Runtime Authority Boundary Between LLM Cognition and Autonomous Execution
|
|
0
|
62
|
September 22, 2026
|
|
Is 1.9 Billion Tokens in a Day Normal Yet?
|
|
2
|
190
|
September 21, 2026
|
|
Open ai safety checks blocking benign file writes to google drive folders
|
|
0
|
173
|
September 13, 2026
|
|
What Can AI Safety Learn From a Boiler Code?
|
|
1
|
98
|
September 12, 2026
|
|
AI Safe Mode: autonomous response technology for criminal threats
|
|
0
|
21
|
September 10, 2026
|
|
Could automatic detection save lives?
|
|
1
|
48
|
September 8, 2026
|
|
Author Mode for Responsible Editing of Difficult Scenes
|
|
1
|
72
|
September 8, 2026
|
|
Architecture Proposal: OpenAI Contributor Network — Verified Contribution Discovery, Reputation and Rewards
|
|
1
|
79
|
September 8, 2026
|
|
Astra - "Chat ended as a precaution"
|
|
2
|
382
|
September 7, 2026
|
|
Users Need to Know When They're Doing Factual Research
|
|
1
|
67
|
September 6, 2026
|
|
What if the person ChatGPT stops talking to is your child?
|
|
0
|
30
|
September 3, 2026
|
|
Feedback: A Safety Translation and Inspection Layer for Small Vibe-Coded AI Applications
|
|
0
|
84
|
August 7, 2026
|
|
Feature Proposal: Human-Approved Work Relay between ChatGPT and Codex
|
|
5
|
313
|
June 1, 2026
|
|
Refusal-only safety may push acute-risk users into private escalation
|
|
2
|
78
|
May 23, 2026
|
|
Codex scheme for operations requiring secret access after "setup" runs
|
|
1
|
282
|
May 20, 2026
|
|
Report on observations of methods for exploiting vulnerabilities
|
|
0
|
61
|
March 19, 2026
|
|
Safety Hotline Messages Triggering Harm Instead of Preventing It
|
|
0
|
47
|
March 15, 2026
|
|
Which are your best settings for the strongest cybersecurity in Europe?
|
|
2
|
192
|
March 14, 2026
|
|
Trust-adaptive AI interaction design
|
|
0
|
46
|
March 9, 2026
|
|
Treat AI Policy Frameworks as Empirically Optimized Systems
|
|
0
|
70
|
March 7, 2026
|
|
Preventing Incentive Leakage in ChatGPT App Integrations
|
|
2
|
227
|
February 12, 2026
|
|
Jailbreak resistance benchmark across 52 recent LLMs (7 levels, redacted outputs)
|
|
0
|
175
|
February 9, 2026
|
|
Clearer Warning Messages When Requests Exceed the Scope of Information OpenAI Can Provide
|
|
7
|
303
|
January 27, 2026
|
|
Anthropomorphism, Cultural Context, and AI Interfaces
|
|
0
|
141
|
January 22, 2026
|
|
GPT-5.2 Fabricated Text After Claiming Direct Source Access (Reliability / Trust Regression)
|
|
1
|
103
|
December 27, 2025
|