10XGenThe 2-Day Intensive →
Free resource · nothing to sign up for

Clone or Camera?

A decision framework for AI-generated leadership content — what a clone should make, what has to be filmed by a human, and a policy your organization can adopt today.

Our position, stated plainly. We build AI clones, and we think most organizations should use them for less than they're being sold. The value isn't volume — it's that the repeatable things stop waiting on one person's calendar, so that person has time for the things only they can do.

The sort

Twenty-one kinds of content, sorted.

Filter by verdict, or read it straight through. Print it and put it in front of whoever approves content at your organization.

When you're unsure

Three questions that settle almost everything.

01

Is it repeatable?

Would you say this same thing to a hundred different people without changing it much? If yes, a clone is a reasonable tool. If it only makes sense said once, to one person, it isn't.

02

Is a specific person's specific moment involved?

A death, a gift, a crisis, a milestone, a mistake. If the content exists because something happened to someone in particular, presence is the message. Pick up a camera.

03

Would you be fine if they knew?

Imagine the recipient learning it was AI-assisted. If that's a shrug, proceed. If it would embarrass you or hurt them, that instinct is correct and you should trust it.

If any of the three gives you pause, film it. The cost of filming something you didn't have to is an hour. The cost of generating something you shouldn't have is trust, and trust doesn't come back on the same timeline.

Adopt this

A policy your organization can put in place this week.

Six clauses. Written to be pasted into a staff handbook, a board packet, or a vendor agreement — with no mention of us. It's yours; change anything that doesn't fit.

AI likeness & voice policy — internal
AI LIKENESS AND VOICE POLICY

1. CONSENT AND OWNERSHIP
   We create synthetic likeness or voice models only of individuals
   who have given written, informed, revocable consent. Each person
   retains all rights to their own likeness, voice and name. No model
   of any person is created, used or retained without that consent.

2. SCOPE OF USE
   Synthetic likeness and voice may be used only for repeatable,
   informational content: educational material, answers to frequently asked
   questions, program and product explanations, and adaptations of
   material the individual has already approved in another form.

3. PROHIBITED USES
   Synthetic likeness and voice will not be used for:
   condolences, bereavement or illness; crisis communication,
   apology or correction; personal thanks to a named individual;
   news affecting employment or an individual's standing;
   any statement of legal, regulatory or financial consequence;
   endorsements of a person; or any communication responding
   to a specific individual's specific circumstances.

4. HUMAN REVIEW BEFORE PUBLICATION
   No synthetic output is published, sent or distributed without
   review and approval by a named human being. Approval is recorded.
   Automated publishing of synthetic content is not permitted.

5. DISCLOSURE
   Where content is materially generated using a synthetic likeness
   or voice, we disclose it. Disclosure is plain, visible, and does
   not require the audience to go looking for it.

6. REVOCATION AND DELETION
   Any individual may revoke consent at any time, in writing, without
   giving a reason. On revocation we cease generating new content from
   their model and delete the underlying model within 30 days. Material
   already produced and approved may be retired at their request.

Adopted: ____________    Reviewed annually by: ____________
Disclosure lines — pick the one that fits
ON VIDEO (lower third or first frame)
   Portions of this video use an AI likeness of [Name],
   created with their consent and reviewed before publication.

IN A CAPTION OR POST
   Made with an AI version of me, trained on my own material
   and approved by me before posting.

IN AN EMAIL FOOTER
   Some of the content in this message was produced using an AI
   likeness of [Name], with consent, and reviewed by a person
   before sending.

ON A PAGE THAT USES IT REGULARLY
   How we use AI: we use an AI likeness of [Name] to produce
   explanatory and educational content, with consent. Anything
   personal, time-sensitive, or addressed to you individually
   is filmed or written by a human. We review everything before
   it goes out.
The workflow

Five steps, so approval doesn't become a bottleneck of its own.

1
Sort before you produce
Run the piece through the three questions above at the brief stage, not after it's made. Sorting late is how organizations end up rationalizing something they already paid for.
2
Name one approver, not a committee
One person who knows the voice well enough to notice when it's off. A committee produces delay and diffuses responsibility, which is the opposite of what this needs.
3
Approve the words, then the render
Read the script before anything is generated. Almost every problem is a script problem, and catching it there costs a minute instead of a re-render.
4
Keep a record of what was approved
A dated list of approved outputs. Boring, and the thing you'll want if anyone ever asks — a funder, a journalist, a board member.
5
Review the sort every quarter
What belongs in each column shifts as the technology and your audience's expectations change. Fifteen minutes, four times a year.

Where does your message actually stand?

A clone only multiplies what you've already settled. If you're not sure the message underneath is ready to be repeated at volume, the assessment takes three minutes and shows you which of three constraints is holding you — no signup to see your result.

Take the assessment → About the 2-Day Intensive