Who's all checked out the "generative AI" draft guidelines from China? Any good takes? Going through it today. 🤓
http://www.cac.gov.cn/2023-04/11/c_1682854275475410.htm
@mmitchell_ai The "no false information" and "true and accurate" provisions pretty much rules out all generative AI models, don't you think?
@twitskeptic @mmitchell_ai that all depends on how you define true and accurate. denying uighur genocide is easily covered by the official truth, and admitting it is certainly not officially accurate.
@twitskeptic @mmitchell_ai (i'm quite sure the primary problem totalitarians see in the llms is that the llms produce unpredictable lies)
@mawhrin @mmitchell_ai Exactly. They also assign responsibility to the service providers for generated content which obviously is intended to discourage companies there.
