Real Alpha
Try Pro free

Newcomer Pod · 2 min · Apr 20, 2026

Public constitutions make subjective model grading contestable

Amanda Askell · AI Safety Researcher, Anthropic

Amanda Askell says graders can be checked against human-ranked examples, while publishing a constitution makes the values behind judgment calls visible.

In this episode

Full analysis
  1. 15:03Why AI systems need judgment, not total obedience4 min
  2. 24:24AI self-reports are weak evidence of consciousness5 min
  3. 29:12Uncertainty about AI consciousness calls for basic respectPro pick 3 min
  4. 47:46Verified work contexts can unlock dual-use assistance5 min
  5. 39:00Constitution training draws on existing model knowledge3 min
  6. 31:55AI could expand scarce expertise for rare diseases4 min
  7. 9:23Public constitutions make subjective model grading contestable2 min
  8. 5:00Claude once ended a late-night task like a human colleague3 min
“And then the nice thing about the Constitution being at least out in the world is when you are making judgement calls, you're at least being transparent about it.”
Amanda Askellat 10:48 · Newcomer Pod
10:48

Takeaways

  1. Check graders against examples people have ranked and explained.
  2. Treat judgment-based scores as imperfect.
  3. Publishing the constitution lets outsiders challenge its choices.

About this segment

Amanda Askell says graders can be checked against human-ranked examples, while publishing a constitution makes the values behind judgment calls visible. That gives outsiders something concrete to question.

Amanda Askell

Amanda Askell is an AI safety researcher at Anthropic.

Leads Anthropic's personality alignment team, shaping Claude's character; co-authored the GPT-3 paper at OpenAI.

  • EN

    Eric Newcomer, Host · Newcomer

From “Amanda Askell on AI Consciousness, Claude & Silicon Valley’s Biggest Fear”, Apr 20, 2026. Watch the full episode

Transcript

From the transcript Newcomer Pod publishes with the episode.

  1. 9:27

    adherence the constitution. Yeah. We had one set up where we had made kind of like graders and just looked at the like how much is the model can like behaving in a way that's like consistent with the constitution relative to. It feels like an impossible task to grade, it's such a subjective. Oh yeah, no, it's very hard. I was like, kind of like for a long time, like, you know, because people often know I'm, it's funny because like I love evals and I'm like, if you can find a good way to evaluate a thing, it's, it's really great because you can, you know, you need to be able to tell that something is getting better. And yeah, if you look at like this approach of having models use good judgement and it's, I actually think the same problem exists elsewhere with like tasks that are just bit hard to give a very like concrete score to, you

  2. 10:17

    know, like how good was this poem? You want models to like get better and do well in these things. And actually, it's very like this feels like the, the kind of frontier of like difficulty, you know, rather than these very hard, like but scorable coding tasks. It's kind of things like writing good poetry. And if you took a survey, it could potentially be worse than it was. You know, different expert poets probably have totally different sensibilities. You can't just ask 2 great poets to sort of score it. They might have different.

  3. 10:46

    Yeah, yeah. And I think, and, yeah, some of these things involve judgement calls. And then the nice thing about the Constitution being at least out in the world is when you are making judgement calls, you're at least being transparent about it. And people can give you feedback and you can get a sense of, you know, so if people are like, this seems like a mistake or here's a gap, so they can only see the judgement calls you're making. Yeah.

  4. 11:10

    And I think with the grading, I still think it's very hard. I think the thing you can do is, you know, you can maybe a little bit too in the weeds, but you know, you can take samples where you have a sense of how you would rank them and like why and check that Any kind of like point wise grader that you use to try and evaluate like at least conforms to like, you know, the judgement of people on those rankings.

  5. 11:32

    It's not perfect, but I think that they actually were tracking roughly the thing that we were kind of interested in.