AI Labels Can Reduce Social Media Engagement, Study Finds

A new study found AI-generated labels can reduce likes and engagement, but explaining human involvement may help maintain audience trust.

It feels like every conversation around technology in 2026 involves artificial intelligence to some degree. AI is here. It may be here to stay, but plenty are pushing back, wanting to ensure that human-created content never takes a back seat. A new study shows that people are, in fact, less likely to engage with a social media post once it is made clear that it was created with AI.

The data comes from Science Says, where researchers conducted 8 experiments and analyzed over 1 million TikTok posts. According to their findings, once people knew a post was AI-generated, engagement dropped across the board. Those posts received up to 8% fewer likes and about 7% less overall interaction, including comments and shares, even when the content quality was identical.

Additionally, creators were perceived as putting in 15.6% less effort, and audiences felt about 14.5% less connected to them. There is less stigma surrounding AI when it is viewed as a tool in a more complex process, such as Photoshop’s neural filters.

In video games, the same stigma is found. Recently, the CEO of the games suggested that Valve should not require developers to disclose AI in development, as it is a Scarlet Letter for developers and forces them to choose between longer development times and the risk of unintentionally upsetting the player base. Read those full comments here.

How do you feel about this data’s findings? Sound off about this story and all others in our official Discord server.

For more AI discussion, check out what Sony has to say about its AI usage going forward. Also, don’t forget to sign up for our weekly newsletter.