Are actors’ agents starting to feel differently about AI?

Something has shifted in the way actors’ agents are talking about AI.

I can see hear it in our conversations.

In late 2024 and early 2025, I ran various brainstorming sessions and workshops on AI for agency teams, from small independent agencies to some of the larger ones. The rooms were full, so clearly the interest was there, but so was the concern.

At the time, the dominant sentiment in the room was often something like (and I’m paraphrasing):

“This is important. We need to understand it. But ultimately, we’re waiting for a union deal.”

A sort of “let’s wait and see”, and in some cases, the feeling was closer to “let’s wait …to be saved…by a union deal”.

A sentiment I completely understood at the time, but did not agree with then, nor do I now, as I’ve explained here.

Fast-forward a year or so and the questions I’m getting are very noticeably different. The energy within the agent community has shifted towards:

What can we do now?

What can we put in our contracts?

What should we ask before our client accepts the job?

Should our agency have a position on AI?

Do we tell clients and buyers what that position is?

Should be loud or quiet about it?

And, increasingly: where are the opportunities?

Quite the shift in less than 18 months.

My AI workshops, webinars and Bootcamp started filling up with agents asking increasingly practical questions.

I wish I could tell you this is because my marketing suddenly became brilliant. It isn’t. (I do very little marketing beyond emailing my list when I’m running something. Something you already know if you’re reading this, as you’re getting those emails in bursts!)

The mindset of agents changed, as AI became more and more discussed in deal breakdowns and starting, showing up in contracts across different sectors of the acting industry. “AI pilots” were being pitched to talent agencies, large and small.

Some of those opportunities looked interesting, most did not. Either way, the question is the same: do you share it with your clients, if so how do you advise your client on whether it’s an opportunity to decline or explore a little further?

Simply put, how do you become a “good” agent – diligent and aware – for AI deals? …even when there is no union deal yet.

The mindset shift I sensed in agents this year did not happen in a vacuum. You can pinpoint the twists and turns the chatter on AI coming from the agents’ community to key milestones and changes brought by AI deployment in the acting community. In other words, because the market changed, the mindset of agents did as well. 

I’ve been working on digital imitations of performers (what we would now probably call digital replicas or deepfakes) since 2013, when I started my PhD on the topic.

So I’ve watched both the technology and the industry’s reaction to it for quite a long time.

When commercial AI voice models started becoming seriously good (say, around 2020-2021), the first response I noticed amongst voice agents was concern, with a drop of fear.

Where does this leave our artists? Where does that leave as their agents? Is AI going to replace voice work? Is it going to create new work? And what on earth are we supposed to do about it?

Then SAG-AFTRA went on strike in 2023, and AI became one of the major issues on the negotiating table.

In the UK, attention naturally turned to Equity and PACT.

And for quite a while, the response I heard from many agents was: we need a collective agreement.

They weren’t wrong.

One agent negotiating one contract for one actor cannot solve an industry-wide problem. Collective bargaining matters precisely because some problems need collective leverage.

But the longer agents waited for UK-wide protections on AI, the more obvious another problem became:

Deals weren’t waiting. And what industry standards are we risking to set in the interim?

AI language was already appearing in contracts. Actors were already being asked to agree to scanning, digital cloning, AI training and other uses. Agents still had contracts to review, clients to advise and deals to negotiate.

So some started working out what they could do during this ‘interim’ period.

I saw this particularly early amongst voice agents.

• Some started adding AI rights reservations to their emails (here is an example).

• Some asked questions about AI before allowing a job to progress.

• Some had their clauses and contracts reviewed by senior peers or lawyers.

• Some started pushing back on vague permissions and broad rights transfers that could leave the door open to uses nobody had actually discussed.

They were trying to strike a careful balance: protect the client from broad or unclear AI permissions without killing perfectly good work in the process.

And, at the time, it could feel a little lonely.

If you’re the agent asking the “AI question” when nobody else seems to be doing it, or redlining the clause everyone else appears to be accepting, you can feel as though you’re sticking your head above the parapet.

You might wonder, am I protecting my client here, or am I about to lose them the job?

That’s not an academic question when your client needs to pay their rent or school fees with the revenue from that job.

But neither is the other side of it: what happens if their voice, face or performance can be reproduced in ways neither of you properly understood when the contract was signed?

That tension around AI hasn’t disappeared. What’s changing is agents’ response to it.

The questions are no longer only about keeping AI out.

Increasingly, agents are asking what it might look like to let AI in selectively, deliberately and on the right terms.

That shift came through very clearly in a recent conversation I had with a senior actor agent specialising in voice work. I’ll call her Sam for the sake of putting her point of across discreetly.

Sam described moving from steering clear of AI jobs to having artists saying, essentially:

“I can see it’s here to stay, so let’s look at what the opportunity might be. My clients will probably turn it down at the end of the day, but even then I want to be able to say I’ve considered all our options. And maybe one day, the right deal on the right terms will turn up, and we’ll say ‘let’s go’.”.

That doesn’t mean everyone suddenly loves AI.

Far from it.

I see agents and actors who want nothing to do with it. Others who are cautious but open to the right project. Others who are actively curious about new commercial opportunities.

And increasingly, I’m seeing the same questions emerging at agency level.

What is our position on AI? Do we even need one?

Do we openly say we’re an agency that takes AI protection seriously?

Do we position ourselves for or against certain uses of AI and make that part of what we stand for?

Do we actively seek responsible AI opportunities for clients who are interested?

Do we say nothing publicly but develop clear internal rules for what we will and won’t accept?

Or do we wait and see?

And I think that’s the real shift I’ve seen between 2024 and 2026.

It is not that agents have stopped worrying about AI entirely. 

There’s still an element of fear. And a degree of frustration at the situation their clients find themselves in (each day a bit more squeezed on rates and working conditions).

And for that, there is also a healthy dose of anger and some scepticism (entirely justified scepticism given the amount of hype and scaremongering around AI we’re bombarded with in the press and social media).

But there’s also more curiosity. More pragmatism. And, importantly, more of a sense of agency.

The conversation has moved beyond “when is somebody (Equity?) going to sort this out?”

to

What can we do in addition to, around and while we wait for those protections?

And that’s a much more useful place to be.

Because there are things agents can do.

• You can ask what AI use is actually proposed before the job is agreed.

• You can identify where an ordinary rights clause might inadvertently capture AI uses.

• You can separate permissions for the original performance from permissions to create or use a digital replica.

• You can ask for informed consent, limits, control and compensation.

You can decide what your agency’s baseline is, and where you’re prepared to negotiate.

And you can start having a proper conversation with your clients about what they want.

Not every client will give you the same answer. Much like, not every agency should have the same position either.

So, where are you?

Where are you in your journey with AI?

Where is the rest of your team? Are your junior agents in the same place as your senior agents? What about your clients?

And perhaps the bigger question:

Are you still waiting to find out what the industry’s position on AI will be, or are you starting to work out what yours is?

Yes, that’s a lot of questions, I’m firing at you right now. I’d love to know if you recognise yourself in any of the above.

If you feel like sharing, drop me a line at mathilde@agentscounsel.com.

If you feel like joining the (growing!) community of agents keen to get clear on AI to make it work for them and their clients, join one of our live event coming up this autumn. I’ll be running free live webinars and my AI Bootcamp for Agents.

Free webinar on ‘Union deals for actors on AI: where are we now?’ (8 October 2 pm (UK).) — Click here to register.

Here are some additional free resources that might help you in the meantime:

• A checklist of questions to ask when negotiating an AI voice deal, or project involving a digital replica (visual).

• An template for an AI protection clause to add to your email signature to reserve AI permissions from deals negotiated via email.