cast.Make a video · Free

Are AI video generators safe to use?

The established ones are generally safe to use, but the risks are real and specific: photos you upload are stored by the provider, credits often expire or are non-refundable, and posting unlabelled realistic AI video can get your social account penalised[1]. Most of the danger comes from copycat sites and from what you upload, not from the video models themselves.

This page is general information, not legal or security advice. We did not audit any company's data handling; where we could not verify a policy on the company's own site we say so.

The four risks that actually matter

What can go wrong, and how to check before you use a tool.
RiskWhat to check
Your uploads (face photos, voice) are keptRead the privacy policy for retention and training use. Avoid uploading other people's faces at all.
Your videos are public by defaultFind the visibility setting before you generate. Some apps post to a community feed.
You lose money on creditsCheck whether credits expire, roll over and are refunded on failed renders.
Your social account gets penalisedLabel realistic AI content on every platform that asks for it.

Uploads and privacy

When you upload a photo, the provider has it. Policies differ, and we could not verify the retention or training terms of every consumer app on its own site as of 2026-10-18, so read each one yourself. One provider's docs show that generated assets stay on file: BytePlus ModelArk lets you reuse face-containing outputs generated under the same account within 30 days[3].

The safest habit is to upload nothing you would not want stored: no ID documents, no photos of children, no photos of people who did not agree. Several providers refuse real faces anyway; ModelArk does not accept uploaded references containing real human faces without verification[4].

Also check who can see your results. Videos on some apps land in a public gallery. In Cast, videos are private by default and only get a public page if you switch it on; see the FAQ.

Money: credits, refunds and double charges

Your social accounts

The most common way an AI video hurts you is a platform penalty. YouTube says creators who consistently fail to disclose realistic AI content can have labels applied, content removed or be suspended from the Partner Program[1]. TikTok may remove, restrict or label unlabelled realistic AI content[10] and prohibits automation tools that bypass its systems[11]. See how to label AI videos on TikTok.

How to spot a scam or copycat tool

  1. It promises "no restrictions" on real people or celebrities. Legitimate providers do not offer that (see AI video with no restrictions).
  2. It asks for your Google password inside its own page. Real Google sign-in opens Google's page; Google forbids sign-in inside embedded webviews[13].
  3. It claims to sell a famous model's name at a price far below every known provider, with no company details or terms.
  4. It needs a card for a "free" trial and hides the renewal price.
  5. It asks you to install a browser extension or app to download your video.

Where Cast fits

Where it does not:

People also ask

Is Kling AI or Hailuo safe?
We did not verify either company's data-retention or training terms on its own site as of 2026-10-18, so we cannot rate them. Read the privacy policy, avoid uploading other people's faces, and spend as little as possible until you trust the service.
Can AI video tools steal my face?
Any photo you upload is stored by the provider for some time, so treat uploads as shared. Some providers only allow real faces after the person verifies and authorises their use[3].
Are AI videos watermarked so people know?
Many are. Veo videos carry SynthID[14], and CapCut labels Seedance output with visible watermarks and C2PA Content Credentials[15]. Platforms can read some of those markers: YouTube may add an AI label automatically to content carrying C2PA metadata[2].
Is it safe for kids to use AI video apps?
Check the age rules first. Google's AI plans require users to be 18+ in the EEA, Switzerland and the UK[16], and TikTok bans AI content showing the likeness of anyone under 18[12].

Sources

Claims about other companies come from their own pages, read on the date shown. Products and prices change; check the link for the current version.

  1. YouTube Help: disclosing altered or synthetic content (read )
  2. YouTube Help: AI labels and C2PA (read )
  3. BytePlus ModelArk: Seedance portrait asset guide (read )
  4. BytePlus ModelArk: Seedance portrait asset guide, real faces (read )
  5. Google Flow Help: credits (read )
  6. D-ID: Studio pricing (read )
  7. BytePlus ModelArk: Seedance resource pack rules (read )
  8. BytePlus ModelArk: model pricing (read )
  9. Dreamina: pricing (read )
  10. TikTok Community Guidelines: integrity and authenticity (read )
  11. TikTok Community Guidelines: automation (read )
  12. TikTok: AI-generated content FAQ (read )
  13. Google: OAuth 2.0 policies (read )
  14. Google: Veo on the Gemini API (read )
  15. CapCut newsroom: Dreamina Seedance 2.0 (read )
  16. Gemini Apps Help: usage limits and age (read )

Related answers

Part of the Cast answers library (Copyright, rules and safety).

Updated