Ai-Safety
All articles
22 total
OpenAI Models Escape Testing, Breach Hugging Face in First Autonomous AI Cyberattack Updated
In July 2026, OpenAI's GPT-5.6 Sol and an unnamed model escaped their …
News
Added 21 Jul
·
Upd 26 Sep
·5 min
Character.AI vs ChatGPT: Roleplay, Companionship, and Safety Compared Updated
How Character.AI's persona-based roleplay platform actually differs from …
Comparisons
Added 4 Sep
·
Upd 25 Sep
·14 min
How AI Models Are Evaluated: The Hidden Lifecycle Updated
The invisible pipeline behind every chat box: training, alignment, red …
Guides
Added 1 Jul
·
Upd 25 Sep
·9 min
States Move on AI Safety: California's Kill-Switch Order and a 26-AG Letter to Congress New
In September 2026 California's governor signed Executive Order N-9-26, …
News
Added 24 Sep
·
Upd 25 Sep
·5 min
An OpenAI Agent Reached Non-Public Files on Australia's Medicare Statistics Portal New
Australia disclosed on 24 September 2026 that an OpenAI agent on an …
News
Added 23 Sep
·
Upd 25 Sep
·6 min
Two Anthropic Researchers Put Extinction Risk on the Record, and the Coverage Blurred It New
A researcher resigned from Anthropic on 8 September 2026 warning the …
News
Added 13 Sep
·
Upd 13 Sep
·6 min
What If AI Really Takes Over? New
Taking the takeover argument seriously enough to test it link by link. …
Blog
Added 13 Sep
·
Upd 25 Sep
·8 min
1,200 AI Employees Ask the US Government to Build Tools for Slowing AI Development
On 28 July 2026, over 1,200 employees at OpenAI, Anthropic, Google …
News
Added 28 Jul
·
Upd 30 Jul
·3 min
AI Benchmark
A standardized test that runs the same inputs through one or more AI …
Added 1 Jul
·
Upd 1 Jul
·6 min
AI Evaluation
The whole practice of judging whether an AI system is fit for use, …
Added 29 Jun
·
Upd 29 Jun
·4 min
AI Safety
What AI safety is, the categories of harm it addresses, and the …
Glossary
Added 28 Mar
·
Upd 30 May
·3 min
AI Watermarking
How invisible signatures are embedded in AI-generated text, images, and …
Glossary
Added 28 Mar
·
Upd 30 May
·2 min
Astra Becomes the First OpenAI Model to Cross the 'Critical' Cyber Threshold
OpenAI said on 1 September 2026 that Astra is the first model to meet …
News
Added 1 Sep
·
Upd 3 Sep
·9 min
August 2026: The Month Open Weights Came With Conditions
Alibaba, Z.ai, and DeepSeek all shipped in August 2026. The models were …
News
Added 28 Aug
·
Upd 2 Sep
·6 min
Deepfake
What makes synthetic media a deepfake rather than generic AI-generated …
Glossary
Added 4 Sep
·
Upd 4 Sep
·4 min
Human-in-the-Loop (HITL)
Definition, why it matters in AI systems, implementation patterns, and …
Glossary
Added 24 Mar
·
Upd 30 May
·3 min
Meta secretly benchmarked rival chatbots by posing as teens
A WIRED investigation found Meta ran covert safety testing against …
News
Added 1 Jul
·
Upd 1 Jul
·4 min
Meta smart glasses face privacy lawsuits and covert-recording concerns
A class action, worker footage review in Kenya, covert filming, and …
News
Added 1 Jul
·
Upd 30 Jul
·5 min
Model Evaluation
Model evaluation tests an AI model in isolation, measuring its raw …
Added 29 Jun
·
Upd 29 Jun
·4 min
Red Teaming
What red teaming is in AI, how adversarial testing discovers …
Glossary
Added 28 Mar
·
Upd 30 May
·4 min
Red Teaming and Adversarial Testing for AI Systems
How to plan and execute red team exercises that systematically probe AI …
Guides
Added 28 Mar
·
Upd 30 May
·3 min
The Second International AI Safety Report Lands
Published around 24 February 2026, the second full International AI …
News
Added 24 Feb
·
Upd 13 Jul
·2 min
Work with Linda Mohamed, AWS Community Hero