This paper aims to build an actionable framework for permissible online content moderation to combat misinformation. Often strong content moderation policies are invoked when misinformation causes harm. By adopting Mill's ethical framework, I show the complexities involved in permissible content moderation. The conclusion will be that, besides invoking the notion of harm, we should also introduce the idea of cognitive autonomy and adopt useful tools, such as cognitive nudging, to promote a healthier epistemic environment online.