Writing about AI leans on slippery vocabulary — rogue AI, loss of control, existential risk. We annotate documents so the loaded language is visible. Translate your own document →
A demonstration passage written in the style of AI-risk op-eds, annotated by both layers.
15 dictionary 8 model flagsThere is no consensus on the definition of AGI, though OpenAI defines it as AI that outperforms humans at all economically valuable tasks. Often referred to interchangeably with...
An emerging field of research and community dedicated to preventing what they see as AGI's existential threat to human life. Largely computational in focus, the research from th...
Within AI safety, alignment is often used to describe technical means of making models abide by human values. Yet most alignment research does not address what those values are ...
As large language models such as those underpinning ChatGPT were scaled up, they developed capabilities their creators did not anticipate such as the ability to do arithmetic. '...
A phrase derived from long-termism, existential risk refers to threats to human life that would kill the vast majority of the population and leave limited or no prospects of rep...
EA and related communities often appeal to notions of 'human flourishing' and the 'benefit of humanity' while espousing views that are markedly anti-human. By referring to 'huma...
Refers to the thought experiment of an AI system that becomes so rife with capabilities, even if a human being hit the off switch, the system would switch itself back on and pot...
A subfield helmed by podcaster and provocateur Tyler Cowen, closely affiliated with the EA and AI safety movements. Cowen argued that the United States in particular needs a 'sc...