部分数学家担忧 AI 实验室对模型运作的保密增加无谓错误并制造进步假象
Some mathematicians worry that AI labs’ secrecy around their models’ workings increases the chance of needless errors and risks creating an illusion...
AI 导读
部分数学家担忧,AI 实验室对模型内部运作方式过度保密,会提高出现无谓错误的概率,并可能制造出虚假的进步假象。保密文化由此被认为给 AI 研究带来了额外风险。
来源:@TheEconomist · x.com