At MPTS 2026, I grabbed two brilliant minds to talk AI adoption in broadcast. What they said confirmed my hunch: we’re still circling the edges. But the IP land-grab is coming.

By Ben Anchor — Wednesday, 27 May 2026 · Listen to the podcast episode

I’ve been banging on about AI in broadcast for two years now, and at this year’s Media Production and Technology Show in Olympia, I wanted to test a theory: are broadcasters actually building anything in AI, or are they just bolting OpenAI wrappers onto legacy kit and calling it innovation?

So I cornered two people I trust — Vinay Gupta, a broadcast technologist I’ve worked with for years, and Shelly Chambers, a growth marketer and podcast strategist I met on a webinar a few months back. We recorded a quick field episode of the podcast right there on the show floor, and what they said confirmed my hunch: we’re still in the observation phase. But the next act is going to be fierce.

We’re Still Watching From the Sidelines

Vinay put it plainly: “We are still at a very early adoption stage. The broadcast industry in general has been like this. We observe, we watch carefully, and then we see if it makes sense to us now to start heavily investing into this stuff.”

He’s right. And I’ve seen this pattern before — it’s exactly how the industry handled cloud adoption a decade ago. We spent years kicking the tyres, running pilots, wringing our hands over sovereignty and cost. Then, once a few big players jumped, everyone piled in and the whole landscape transformed overnight.

But here’s the rub: AI isn’t cloud. Cloud was about infrastructure — plumbing, essentially. AI is about intelligence. It’s about who owns the decision-making layer. And right now, most broadcasters are renting that intelligence from frontier models rather than building it themselves.

Vinay’s view is that we’re currently in a phase where broadcasters are “leveraging AI by the virtue of keep using these product suites” — essentially, vendor lock-in with a ChatGPT plugin. It’s pragmatic, it’s safe, and it’s deeply uninspiring. The real shift, he argues, will come when broadcasters start building their own custom models on top of their proprietary data — viewer behaviour, scheduling telemetry, archive metadata. That’s where the competitive moat gets dug.

I agree with him on the direction, but I’m less patient about the timeline. He reckons it’ll take “another few years” before we see serious IP development in this space. I think the window is narrower than that. The differentiator, as he put it, will be “companies who have the data strategy in place and very clear understanding of how their data is classified and what is allowed, not allowed to be used for AI.”

That’s the crux. If you don’t have a data strategy now, you’re already two years behind. And if you’re still treating AI as a marketing checkbox rather than a strategic imperative, you’re going to wake up in 2028 wondering why your competitors are making scheduling, advertising, and content decisions three times faster than you.

The Use Cases Are Obvious — But We’re Overthinking Them

I asked Vinay where he’s seen AI applied well, and his answer was refreshingly pragmatic: “Wherever the data is used to make decisions.” Scheduling, targeted advertising, content recommendation — all the workflows that currently involve spreadsheets, gut instinct, and too many meetings.

But here’s where it gets interesting. He also stressed that automation needs to be “with a pinch of human element there” — not “dumbed down workflow, but something which can almost mimic human intelligence in decision-making.”

I walked past three stands at MPTS that were all singing the same hymn: AI-assisted metadata tagging, AI-driven monitoring, AI-optimised scheduling — all with “human in the loop” as the safety blanket. And I get it. In an industry built on editorial control and regulatory accountability, you can’t just let an agent rip. Not yet.

But Vinay went further. He said there will come a time — “not very far future” — when AI agents, given the right guardrails, will make better decisions than humans in certain domains. And when that happens, the question won’t be “should we let AI decide?”, it’ll be “what does the human role even look like?”

That’s the uncomfortable bit. We’re a creative industry. We take pride in craft, in taste, in judgement. But if an AI can schedule a channel more efficiently, optimise ad yield more accurately, and surface archive content more intelligently than a human scheduler, what’s left for the human to do? Vinay thinks that shift is still distant, especially in broadcast. I’m not so sure. I think it’s closer than we’d like to admit, and the sooner we stop treating AI as a co-pilot and start thinking about it as a decision-maker with oversight, the better positioned we’ll be.

Creators Are Ahead of Us — And They Don’t Even Realise It

Shelly’s perspective was a useful counterweight. She came to MPTS to explore AI in creator content-making, and she spent time at the Creators Hub listening to a talk on how to build a podcast that actually retains listeners. The punchline? The creator basically said: forget the shiny production. Audiences don’t care. They want story, consistency, and authenticity.

That’s a lesson broadcast could stand to learn. We’ve spent decades optimising for technical excellence — 4K, HDR, Dolby Atmos, pristine colour grading. Meanwhile, YouTube creators are building eight-figure audiences with a Ring Light and a Rode mic. The gap isn’t technical. It’s cultural.

Shelly uses AI in her growth marketing for podcasts — scripting support, social cards, promotional copy. She’s not using it to make the content; she’s using it to amplify it. That’s the pragmatic middle ground, and it’s where most of us will live for the next few years. But the question I keep coming back to is: when does AI stop being a tool and start being a collaborator?

She made a point that stuck with me: “Everyone, absolutely everyone, is going to have to be able to prompt.” A decade ago, being “good at Excel” was a skill. Now it’s just assumed. Prompting will be the same. If you can’t articulate what you want from an AI in a way that gets you a useful result, you’re going to be at a disadvantage. And that’s a skill we’re not teaching yet.

So Where Does That Leave Us?

Vinay’s right that we’re in the early adoption phase, and that the real value will come when broadcasters start building proprietary intelligence on top of their data. Shelly’s right that the creator economy is already iterating faster than broadcast, and that AI is being used pragmatically in marketing and ops, not just in the obvious content workflows.

But I think we’re underestimating the speed at which this is going to tip. The cloud comparison is useful, but it’s also misleading. Cloud took years because it was a wholesale infrastructure migration. AI is different. It’s software. It’s iterative. And once a few broadcasters start seeing real ROI from custom models — better yield, faster turnaround, smarter programming — everyone else will scramble to catch up.

The question isn’t whether AI will transform broadcast. It’s whether you’ll be building it or buying it. And if you’re still figuring out your data strategy, you’re already late.