Ethics & Society

LLM jailbreaking is the practice of crafting inputs that trick a language model into ignoring its safety rules and producing conte…

LLM security is the practice of protecting AI systems built on language models from misuse, manipulation, and data leaks. Because…

LLM psychosis — often called AI psychosis — is an informal term for cases where intense, prolonged interaction with a chatbot appe…

RAG poisoning is when someone deliberately sneaks false or malicious documents into the knowledge base a RAG system searches, so t…

Bias in an LLM means its outputs reflect the same skewed patterns baked into the text it learned from. An LLM doesn't invent opini…

No court has settled this yet — training AI on other people's writing, art, and journalism is the subject of dozens of active laws…

Yes, but only when the stakes are high enough that not knowing would change how you react — not every single time someone runs a s…

Training a large language model takes a genuinely large amount of electricity and water. That cost doesn't stop once training ends…

AI safety is the field of research focused on making sure AI systems do what we actually want, not just what we literally told the…

AI coding assistants are safe enough for professional use when you treat their output as a draft, not a finished product. The real…

The main risks of generative AI are that it can be used to deceive and that it can fail in quiet, hard-to-spot ways. The headline…

The main security risks of RAG come from the fact that it feeds untrusted documents straight into a model that can act on them. Th…

Prompt injection is an attack where malicious instructions are smuggled into a model through untrusted input — a user message, a w…