I’m doing my thinking about AI in public. You should hold me to it.
I got into an argument. Then I made it a podcast (not least because most takes about AI are terrible...)
Because everything must be polarised these days, most AI discourse boils down to one of two things.
There are those that tell you artificial intelligence is going to save the world: unlock human potential, cure diseases, fix the broken bits of capitalism.
While others will proclaim ultimate doom: AI will take your job, homogenise culture, hand more power to companies that already have far too much of it.
Of course the truth will lie somewhere between these extremes. As individuals, human beings are very good at getting a sense of the truth by figuring out the extremes of an argument and then, if they have the patience, to do the hard work of wading through the messy middle.
Unfortunately media — particularly algorithmically curated digital media — is optimised to grab and hold your attention instead of lead you closer to truth.
So I’m betting that whatever you think about AI right now is closer to one of these two extremes. Can’t we do better than this, particularly when we’re talking about technology that, rightly or wrongly, is shaping so much about what are we thinking and doing as advertising and media people?
I think we can. So I’ve started a podcast.
Believe it or Not is a show about AI in media and advertising — specifically about the big claims being made about it. Each episode takes one proposition and investigates it properly. We make the strongest version of the case, check it against what’s actually happening, and then give you an honest verdict. No prophecy. No panic. Just two people genuinely trying to figure it out.
As you’ll know from previous Ad-verse Reactions pieces about AI and the questionable claims and practices made by companies like Open AI and Meta, I’m very sceptical about how radically AI is going to change the world as these tech companies would like you to believe. That’s not to say it’s all a load of puff, but there is an overriding narrative about AI that needs correcting.
My co-host is Hamish Nicklin — 25 years in media, former CEO of Dentsu Media UK, chief revenue officer at The Guardian, and senior Google/YouTube sales boss. He now runs an AI consultancy called AgentFlow, so he’s in the trenches building with this technology and betting his business on it.
Basically: Believe it or Not is Hamish and I having an argument. We don’t agree on some fundamental points about AI, as you’ll hear.
I didn’t know Hamish that well before we came up with this idea. But when I wrote something sceptical about AI in a Media Leader column earlier this year, he pushed back — publicly, forcefully, and in good faith. More than that, he was willing to clarify and revise his own thinking out loud, which is no small thing when you’ve staked your professional reputation on a position. That kind of intellectual honesty is rare. It’s also exactly the energy this show needs.
The dynamic is straightforward. Hamish is the evangelist. I’m the sceptic.
I also want to be open about how I'm using AI as a journalist and consultant. Publishers owe readers transparency about how they're using these tools — and there are real risks in outsourcing your thinking to them, particularly when the limits of what's useful aren't always obvious. As I like to say about AI and writing: it's like hair dye. It can plausibly do a job for you, but most people can tell something is going on.
Early subscribers will know that one of the first things I published on Ad-verse Reactions was an explainer about how I used AI for writing. Things have changed a lot since then but I’m certainly not a convert. I’m still a work in progress, as I suspect most of you are, too.
I don’t know exactly where I’ll land on all of this. But I’d rather figure it out in public than pretend I already know.
Episode zero is out now on The Media Leader’s podcast feed. It’s a limited series and full episodes drop every Thursday.
As the series develops, I want to hear what you think — not just whether you agree, but where you think we’re getting it wrong. That’s the whole point. Reach me at omar@oomphoomph.com or find me on LinkedIn.