Then they let natural-language-processing algorithms learn from the real human responses and craft their own, such as: " I don't think using words that are sexist in nature contribute to a productive conversation."

然后，他们让自然语言处理算法学习真人回复，并创作自已的回复，比如：“我认为使用本质上带有性别歧视的词语无助于形成有效对话。”

Then they let natural-language-processing algorithms learn from the real human responses and craft their own, such as: " I don't think using words that are sexist in nature contribute to a productive conversation." Which sounds pretty good.

然后他们让自然语言处理算法从真实的人类反应中学习并制定自己的反应，例如：“我不认为使用本质上带有性别歧视的词语有助于进行富有成效的对话。 ” 这听起来不错。