U.S. consumers struggle to distinguish between videos recorded by humans and those generated by Sora, OpenAI’s text-to-video tool, according to new data from HarrisX provided exclusively to Variety Intelligence Platform (VIP+).

A survey conducted a few weeks after the controversial software was introduced found that most American adults incorrectly guessed whether the creator was AI or a human in five out of eight videos they were shown.

Half of the videos were Sora demonstration videos that went viral online, raising concerns about production quality from Hollywood to Capitol Hill. They include drone footage of waves crashing against the rugged cliffs along Garay Point Beach in Big Sur and historic footage of California during the Gold Rush.

Perhaps unsurprisingly, the HarrisX survey also found that a large majority of respondents believed the U.S. government should implement regulations requiring AI-generated content to be labeled as such. They were equally emphatic about the need for regulations for all content formats, including video, images, text, music, captions, and sounds. The full HarrisX survey results can be found on VIP+.