Seth Godin 谈 AI 内容的恐怖谷不适感为何会随时间消退
The creep fades
- 1作者提出恐怖谷式的不适感会随接触增多而消退，人类历史上多次经历同类过程
- 2文章观察到平庸的 AI 风格文字泛滥后，人类开始下意识模仿这种文风并逐渐接受
- 3作者以爱迪生的活动影像曾令观众尖叫、香菜辣酱需适应为例，说明警觉是演化本能
- 4作者提醒警告消退不等于问题消失，文化中受信任的熟悉元素也在悄然流失
公众对 AI 生成内容的不适感是适应期的过渡现象，频繁曝光本身会推动接受，与内容质量是否过关并无必然关系。
- 影响：受众对 AI 内容的抵触会随曝光自然减弱，初期反感不必直接决定内容的长期取舍；员工看惯平庸 AI 文风后会下意识模仿，品牌语调同质化是经营层面的风险；文中提醒受信任的文化元素正在同步流失。
- 参考：可考虑为品牌语定一份明确的文字规范，把 AI 生成稿与人工终审分开，避免团队文风被默认输出带偏；值得盘点内容资产里依赖受众信任的元素，先判断哪些要保留人工投入。
- 依据：这是 Seth Godin 的个人观点专栏，基于历史类比与文化观察写成，没有数据或独立评测支撑，判断属于作者个人观点。
The uncanny valley is the place where an image that’s almost human creeps us out:
Sorry to share that.
Human beings evolved to have a distaste for things that are almost, but not quite, the usual thing. It takes a while for someone who wasn’t raised with cilantro to like it–the same with hot sauce. When Edison first showed people his moving pictures, people screamed. Our brain is warning us that categories are getting surprising and we should be on alert.
As AI spreads through the culture, we’re seeing plenty of mediocre AI-style writing. And then humans subconsciously start writing in that style because they see it so often. And finally, a grudging acceptance that it sounds okay. We’ve been through this a hundred times before.
No one asks to get creeped out. It’s usually done to us, not for us. But, over time, the creep fades.
Like most warnings, we ought to keep two things in mind:
- All warnings are not the same. We need to decide which ones to ignore.
- Just because a warning fades doesn’t mean that the problem is gone.
The creepiness fades.
But at the same time, there’s a creeping change as trusted and expected elements of our culture fade away.
来源：Seth Godin · seths.blog