OfferNew accounts keep every subscription at 85% of the list price, forever, plus 145 free credits. Extra credit packs include 5% more.

Create account
Scro
Sign in
← Back to Posts
ChristmasAI VideoTalking Head

A Realistic Voice, Say It Is AI

If the voice sounds real, say it is AI. Leave the platform's own label where the app puts it, beside the offer.

See this product
A Realistic Voice, Say It Is AI

If the voice is realistic, say that it is AI. Jun speaks for Hearth Bakery in a voice that sounds like a person, so the caption and the post say the voice is AI. One person speaks. You also say it when the face is realistic. Leave the platform's own disclosure where the app places it. The spoken offer can still be sentence one: they close at four. The disclosure sits with the post. It does not replace the offer, and the offer does not replace the disclosure. The disclosure is a true sentence beside the offer, and it has an order. Approve the still, let Jun say they close at four as sentence one, write in the caption that the voice and the face are AI, and leave the platform label where the app puts it. The first three seconds show him at the counter while he says they close at four. A realistic file whose caption never says the voice is AI, or a disclosure that replaced the offer, is rejected before posting. One speaker, fifteen to forty-five seconds.

Say it in the caption

Write a plain disclosure that the voice is AI, and that the face is AI if the face is realistic too. Put it where viewers see it, with the shop line. Hearth Bakery's offer stays intact: they close at four. The disclosure is an extra true sentence, not a substitute for the hour. Burned-in captions can carry the offer in the upper-middle while the post copy carries the AI label. Write the disclosure next to the shop line, and keep the hour intact. The offer can be burned upper-middle while the post copy says the voice is AI. The first three seconds show the closing time while Jun says they close at four. A caption where the AI line replaced the hour is the version you reject.

Leave the platform label

Apps often add their own mark when a face or voice is realistic. Leave that mark where the app placed it. The job in this reel is to say the voice is AI and to keep speaking as one person, Jun. A cartoon voiceover on a drawing is a different kind of file. This one is realistic, so it is labeled. After the offer is sentence one, write the disclosure. The app's own mark stays where it was placed. One person speaks, and this realistic file is labeled in the caption as well. The opening three seconds show Jun while he says they close at four. A post missing the line that the voice is AI is rejected. The caption still says the voice is AI.

The label does not fix the still

Saying the voice is AI does not excuse a drifted face, indoor snow, or a buried offer. Approve the still first. Then speak. Then label. Export at a practical fifteen to forty-five seconds if this is the counter post. The first three seconds still carry the closing time. The AI line is honesty beside that point, not instead of it. Still, then speech, then label. The AI line does not excuse a bad picture. The first three seconds show a clear mouth while Jun says they close at four. Indoor snow, a drifted face, or a buried offer is rejected before posting, even when the disclosure is ready.

A worked example

Jun at Hearth Bakery says they close at four in a realistic voice. The caption states that the voice and the face are AI. A platform label, if the app adds one, stays on the upload. One person speaks. The offer remains sentence one, with the disclosure written beside it in the post. Still, offer, AI line in the caption, upload. The first three seconds show Jun while he says they close at four. A realistic voice with no AI disclosure is rejected before posting. Any platform label stays on the upload.

Before you post

  • The caption says the voice is AI.
  • Any platform label is left in place.
  • One person speaks the offer.

Where Scro fits

Scro keeps one person speaking and can set the spoken line in one of 13 caption styles. Export at 9:16. Label a realistic face or voice as AI, and leave the platform's own label on the post. Speak the offer, set the line in a caption style, label the realistic voice, and export after the hour is still sentence one. The first three seconds show Jun while he says they close at four. A post whose caption never says the voice is AI is rejected. Leave the platform label where the app puts it.

Label the realistic voice. Finished samples are on the examples page.

Related: Four Beats for One Policy.