Skip to main content
When your videos show a real colleague, you are using something that belongs to them. LatentCast is built so that this is always their decision, and so you can say that plainly to your people and to anyone who asks.

The principle

No colleague’s face or voice is used in a produced video until they have granted consent, in their own account, for that specific outreach. There is no administrator override, and no path through the product that skips it.
It is enforced where it matters, at production. A recipient whose presenter has not consented is not produced, and the batch tells you so rather than quietly substituting somebody else.

What each person controls

  • Whether to take part at all. Declining is a first-class answer.
  • Which likeness and which voice are used. They choose from their own library, not from a pool.
  • Their material. Their photograph and recording live in their account. The colleague running a campaign cannot see or change them.
  • Changing their mind about the setup. A new avatar or voice makes the previous agreement stale and asks again, because they agreed to something specific.
One outreach, the recipients assigned to that person within it, using the avatar and voice they picked, in the languages their voice covers. A second campaign asks again.

What to tell a colleague you are casting

They will want three things answered, so answer them first:
  1. Where does this go? To the named customers on their list, on a private page reached by a link they are sent.
  2. What am I agreeing to? This outreach, this avatar, this voice. Nothing else, and it can be withdrawn by talking to you.
  3. How much of my time? About fifteen minutes, once: a photograph, a short recording, and one confirmation.
You have been cast answers all three, and is written to be sent to them directly.

Recipients, and being honest with them

Every video LatentCast produces is marked as AI-generated: visibly, with a disclosure badge you position yourself, and invisibly, in ways that survive being copied and re-encoded. That is a legal requirement in the European Union and it is also simply the right way to use this technology. AI transparency covers what is marked, how anyone can check a file, and how to read the result.

Next

AI transparency

How our videos are marked, and how to verify one.

Presenters and consent

The practical side: inviting, tracking and chasing.