I hope their detector is better than the typical 'AI detection in text' services. False negatives are bad, false positives are worse as some creators could lose their source of income.
I hope their detector is better than the typical 'AI detection in text' services. False negatives are bad, false positives are worse as some creators could lose their source of income.
At least for now, YouTube honors the uploader’s label of whether or not the content was AI generated. Only when it’s left unset do they do any automatic labeling. And those labels can be overridden manually by the uploader if they get it wrong.