Prakash#1330
"90% reliable, but for a Catholic, these aren't negotiable"
Matthew Sanders, CEO of Catholic AI firm Longbeard, explains why Catholic AI can't just be 'pretty good.'
His argument is that if a model is wrong on the one morally decisive case, that's not acceptable—and that's why steerability and value alignment matter so much.
@SandersGame
I did actually test this, specifically the euthanasia question, and the all the models were pro. Chinese or Western, open source or frontier.
"90% reliable, but for a Catholic, these aren't negotiable"
Matthew Sanders, CEO of Catholic AI firm Longbeard, explains why Catholic AI can't just be 'pretty good.'
His argument is that if a model is wrong on the one morally decisive case, that's not acceptable—and that's why steerability and value alignment matter so much.
@SandersGame
Follow @AI_in_the_AM for the daily rundown!
Full video: https://briefing.ai-in-the-am.com/p/aiam-self-improving-tax-agents-and-catholic-ai
Prakash#1330"90% reliable, but for a Catholic, these aren't negotiable"
Matthew Sanders, CEO of Catholic AI firm Longbeard, explains why Catholic AI can't just be 'pretty good.'
His argument is that if a model is wrong on the one morally decisive case, that's not acceptable—and that's why steerability and value alignment matter so much.
@SandersGame