AI made posting nearly free. Trust is the scarce thing now

On 14 July 2025, Meta published new rules against unoriginal content on Facebook. The next day, YouTube renamed its “repetitious content” rule “inauthentic content”. It made clear the rule covers content that’s repetitive or mass-produced.

Neither banned AI. Both drew a line around effort and originality, at the same moment and for the same reason. A month of social posts that used to take a team weeks now takes an afternoon. Something that costs almost nothing to make stops being a way to stand out.

Every brand can now publish daily in six formats. So the question for marketing leaders is what to do with the hours AI hands back.

What did the platforms rule?

YouTube’s policy page says AI-generated content made from generic or unoriginal templates, that looks mass-produced, can’t earn money. It says such content never could. It still allows AI tools for jobs like script editing, when the finished video shows creative input.

Meta defines unoriginal content as reusing another creator’s work again and again without credit. Accounts that do it lose monetisation for a period and see their reach cut on everything they share. Stitching clips together or adding a watermark doesn’t count as meaningful change.

I read both as a signal of where the algorithms are heading.

What do audiences say?

Sprout Social surveyed 2,250 social media users in the US, UK and Australia in February 2026. Its release reports that:

What consumers most wanted brands to stop was posting AI content without labelling it. Sprout sells social software, and some questions are about news, not brands. So treat the figures as rough signals.

A narrower academic result points the same way. A Washington State University team of Mesut Cicek, Dogan Gursoy and Lu Lu ran experiments with more than 1,000 American adults across eight kinds of product. They published in 2024. Describing a product as using artificial intelligence lowered purchase intent, because it lowered emotional trust. The effect was bigger for risky purchases.

That study is about the words “artificial intelligence” in product descriptions, not AI-made posts. But it should end the idea that “AI-powered” is a free upgrade to any message.

What I’d do with the hours

I led a growth team that used AI for research, reporting and landing pages. The lesson I took: freed time can become more output or better work. Output wins by default because it’s easy to count.

I’d cut the posting plan and keep the team. If a post can’t answer “what does this say that only we could say?”, it doesn’t go out.

Spend the hours on proof. A customer’s real number. A screenshot of the actual result. A chart from your own data. Your listening work will turn these up.

Give the editor a bigger job, because a draft nobody changed reads like everyone else’s. That means voice, facts and one added detail per post.

Decide your disclosure rule before anyone asks. Label where a platform asks, and always for realistic AI images and video. My working rules are in AI workflows for social media teams.

And keep the technology out of the headline, unless your audience has told you the AI part helps.

The case against

Audiences are split. Plenty of people don’t mind AI content if it’s useful. A brand that posts less can lose reach in a feed that rewards frequency.

I can’t prove my view from a survey. But the costs are lopsided. A careful post costs a little reach. A feed that teaches your audience to scroll past you is hard to undo.

So test it for four weeks. Post half as often, with twice the original material in each post. Compare saves, shares, replies and qualified sales conversations against the month before. If reach falls and nothing else rises, I’m wrong for your audience, and you’ve lost four weeks.

Sources