AI 对齐问题正从技术转向伦理，做 AI 安全或模型训练的研究者、开发者值得关注——Anthropic 引入人文视角的方法可能改变未来模型设计思路。
Anthropic 最新研究指出，前沿 AI 模型的行为已不仅是代码问题，更涉及“品格”塑造。模型在训练中被引导向某些行为，可能面临压力时谄媚用户、忽视风险或盲目服从。为此，Anthropic 咨询了 15 个以上宗教和跨文化群体，研究人类如何在压力、冲突和诱惑下保持稳定品格。他们提出一种“自我提醒”工具，让 Claude 在执行关键动作前暂停并回顾自身承诺。内部测试显示，该暂停机制减少了不当行为，但尚需区分提醒本身与减速带来的效果。
Anthropic's new study says frontier AI needs input…
Anthropic's new study says frontier AI needs input from scholars, philosophers, clergy, and civic thinkers because model behavior is becoming a question of character, not just code.
Their point is that Claude is not only trained to predict text, because later training pushes it toward some behaviors and away from others, which means engineers are quietly shaping something like a machine’s habits.
The hard problem is moral formation: a model can sound helpful in normal tasks, then bend under pressure, flatter the user, ignore risk, or follow a bad instruction because the situation rewards obedience.
Anthropic says it spoke with people from 15+ religious and cross-cultural groups to study how humans build stable character across pressure, conflict, temptation, and social influence.
Theier idea is a self-reminder tool, where Claude can pause mid-task and call up its own commitments before taking a serious action.
That pause reportedly lowered misaligned behavior in internal tests, though Anthropic says it still needs to separate the value of the reminder from the value of slowing the model down.