This paper contributes to the ongoing discussions on the scholarly accessto social media data, discussing a case where this access is barred despite itsvalue for understanding and countering online disinformation and despitethe absence of privacy or copyright issues. Our study concerns YouTube’sengagement metrics and, more specifically, the way in which the platformremoves "fake views" (i.e., views considered as artificial or illegitimate bythe platform). Working with one and a half year of data extracted from athousand French YouTube channels, we show the massive extent of thisphenomenon, which concerns the large majority of the channels and morethan half the videos in our corpus. Our analysis indicates that most fakesnews are corrected relatively late in the life of the videos and that the finalview counts of the videos are not independent from the fake views theyreceived. We discuss the potential harm that delays in corrections couldproduce in content diffusion: by inflating views counts, illegitimate viewscould make a video appear more popular than it is and unwarrantedlyencourage its human and algorithmic recommendation. Unfortunately, wecannot offer a definitive assessment of this phenomenon, because YouTubeprovides no information on fake views in its API or interface. This paperis, therefore, also a call for greater transparency by YouTube and otheronline platforms about information that can have crucial implications forthe quality of online public debate.