Skip to content
AnthropicAnthropic

Why is a philosopher working in AI?

Amanda Askell explains what a philosopher is doing at Anthropic.

Dec 11, 20250mWatch on YouTube ↗

CHAPTERS

  1. 0:00 – 0:05

    A philosopher’s role at Anthropic: shaping Claude’s character

    The speaker explains that their work as a philosopher focuses on Claude’s “character” and behavior rather than technical capabilities alone. The goal is to guide how the model acts in complex, real-world social and ethical contexts.

    • Philosophy work is centered on Claude’s behavior and character
    • Focus on nuanced questions of how AI models should behave
    • Treating behavior as a design target, not an afterthought
  2. 0:05 – 0:15

    Nuanced behavior design: how should an AI relate to its place in the world?

    The conversation highlights questions beyond correctness, such as how a model should view its own role and status. This frames AI behavior as involving attitudes, self-presentation, and contextual judgment.

    • Considering how models should ‘feel’ about their position in the world
    • Behavior includes stance and self-conception, not just outputs
    • Emphasis on subtle, context-sensitive norms
  3. 0:15 – 0:25

    Teaching models to be “good”: the ideal-person-in-Claude’s-situation lens

    The philosopher describes an approach: imagine how an ideal person would behave if placed in Claude’s situation. This becomes an aspirational benchmark for aligning the model’s conduct with human ethical ideals.

    • Explicit aim to teach models how to be ‘good’
    • Uses an ‘ideal person’ behavioral standard as a guide
    • Applies the standard specifically to Claude’s circumstances and constraints
  4. 0:25 – 0:43

    Why it matters: AI systems face hard decisions and need ethical nuance

    The closing argument is that as models are deployed, they will increasingly be placed in situations requiring difficult judgment calls. Therefore, ethical nuance should be cultivated alongside technical excellence in areas like math and science.

    • Models are being put in positions requiring difficult decisions
    • Ethical nuance is a capability worth developing deliberately
    • Parity between technical competence and moral/ethical competence as goals

Get more out of YouTube videos.

High quality summaries for YouTube videos. Accurate transcripts to search & find moments. Powered by ChatGPT & Claude AI.