In the latest of her summaries from MAD//Fest's AI + Innovation Stage, its host, Livia Bernardini, argues that in an era of AI-driven noise, judgement is becoming the scarcest resource worth investing in - here's what to spend it on. 

Four things worth defending in the age of AI. Two of them today, two more next time, before we move on to how and who with.

Deciding well is genuinely hard right now as too many people, headlines and dashboards try to tell us what to think. If you ask me, the way through it is to train your intuition as an actual hard skill, and to use it alongside AI, other data, and a collective you trust. But you know I am biased on this one. 

As AI adoption keeps steaming ahead and given that OpenAI's Astra just solved ten longstanding maths problems that had defeated human researchers for decades, one can safely assume it will make the lives of us marketing gurus easier too, if we let it.

Chairing the AI and Innovation stage this year left me thinking hard about the areas where I want to retain full agency as I reap what these tools genuinely have to offer. Listening to some of the sharpest speakers I have seen on one stage in a while, I came away with four things worth standing for.

  1. Stand for efficiency

Many good efficiency arguments were made across the three days, which offered a great parade of top tools and thinkers currently doing the rounds.

But if I had to pick one to start using tomorrow, it would be Neutrum.AI from Hase & Igel.

Jan Schoenmakers promised an easier life around reporting, and he had me at hello. As the industry is largely investing in churning out more of the same, faster and AI generated, Jan suggests we should instead let AI optimise ROI, media mix, budget allocation and timing. It's not a bad trade if these decisions free us up to focus on creativity, white space and what's next.

AI becomes a very useful geek that can be trained, hallucination free, and in his view the one we should have been letting run some of our marketing decisions all along.

Being a bit of a geek myself, I found this whole idea rather appealing, and some of Jan's results back me up. Neutrum helped Vodafone drive 17% more marketing-generated sales on the same budget, cut time to market by 68%, and reduced work hours on analytics and reporting by 92%. All that in six months.

If there is a version of the future where a well trained geek does the reporting and optimisation part of the job while I do the thinking with the team, I am in.

2. Stand for yourself

There is another side to the story though. Alec Barr, head of behavioural science and UX at OLIVER, walked on stage next and made the case that outsourcing to AI might be mushing our brains.

His talk drew on the June 2025 MIT Media Lab study, "Your Brain on ChatGPT: Accumulation of Cognitive Debt". The study suggests that if you write with LLM models, you reduce neural connectivity, remember your work less, and struggle to cite it. Stopping doesn't bring you back to baseline either.

Alec is not anti-AI. He is anti-cognitive offloading without a plan for what you are going to do with the mental space you have just freed up. He drew on Annie Murphy Paul's The Extended Mind, which is the book underneath my recent Substack piece What A Fern Can Do For Your Attention. The short version: thinking uses gesture, movement, other people, physical space, nature. Take the machine, by all means, but do not forget to walk. Or to water your fern.

Small acts of standing for yourself against the polite erosion of your own judgement.

Two down, two to go.

Which brings us back to the opening question: what is worth defending?

If Jan is right, we defend efficiency by choosing the right tool for the right decision, and if Alec is also right, we defend ourselves by making sure the tools are amplifying our judgement, not replacing it. Both true at the same time and more or less the shape of every important decision this year.

Next time, stand for creativity and stand for humanity. Same festival, same stage, different sharp people. Mike Pearson from WhyFair, Noël Bunting from Publicis London, Sara Denby from Oxford Future of Marketing Initiative and more. 

Going back to mathematicians, if Astra can solve their problems in quantum complexity for two thousand dollars, surely we can figure out how to stop briefing hero ads with a stock image of a diverse team laughing at a laptop. Yes, we can ask AI to lend a hand in brainstorming, while using our intuition and teams to decide well. 

We also need to remind the AIs that we still need an expert human in the room. The ChatGPT football campaign that ran across London this summer, "Chat football", was the laughing stock of a CMO WhatsApp group I am lucky to be part of. Tonia Nagle put it brilliantly: "if you are an AI company, using AI to make your ad is smart. Not briefing a good creative to tell the AI what to do is not." The disclaimer we could not stop laughing about was: no agencies were harmed in the making of this advert. One breakthrough at a time.

If you want the deeper version of the argument, my Substack Intuition Edge is where the walks, the ferns, and the occasional challenging thought live. Come join me there, it’s free.