An analysis by the Knight Columbia University suggests that deepfakes are relatively rare and are cheap to reproduce without the use of artificial intelligence (AI).
The report also suggests that concentrating on the demand for misinformation is more effective than attempting to limit the supply of misleading content.
This concurs with the widely held view that fact checking is largely ineffective because people largely share content to signal allegiance to a particular group or cause, rather than because the content is accurate or true.
The prevalence of this behaviour means that efforts to tackle the supply of misinformation are not particularly effective.