This needs to be from third parties otherwise those who control the neural nets and datasets will be able to shroud information as "not generated" when it is clearly astroturfing or manipulated. Then they can throw their hands up and say "must be the algorithm or a bad dataset" for plausible deniability.

The new game is coming or here, and it is misdirecting blame to the "algorithm" when it is an editorialized set of data or filtered for certain aims.

Almost all algorithms and datasets are biased or editorialized in some way, laws need to be adjusted on that. You can't blame the "algorithm" for enragement, because enragement is engagement.