From unoptimised taste and agents as brand activists to one surprisingly reassuring claim about AI and jobs, the AI + Innovation Stage offered some unexpected clues about where human advantage still lies, writes stage host, Livia Bernardini, in the latest of her summaries from MAD//Fest. 

Last week we focused on two things worth standing up for. Number one was efficiency and Jan Schoenmakers made clear that the optimisation of a marketing plan can safely be outsourced (hurray!). Number two was yourself, and Alec Barr reminded us all to approach outsourced thinking to AI with lots of care.  

Today we move up a level, because AI has propelled us into a new order of decision making, one that asks us to answer questions about authorship, originality, craft, ethics and welfare.

Today, two more things worth standing for. 

Number three, creativity, which I firmly believe is the product of the collective. And number four, humanity, because if we’re not building a better future for all of us, then I do wonder what the point of being here is.

So let's dive into it.

Stand for creativity

Stjepan Zelić gave the most interactive, provocative and creatively AI-driven presentation that I saw this year from someone who doesn't hold a Global Creative Director or Chief Strategy Officer title.

Besides setting the record straight about what happened to creativity online post AI, which I covered amply here, and giving us all a little history of art reminder class with his beautifully AI-generated versions of famous works of art in different guises, he made the value of unoptimised taste clear to all of us. In a world where AI has raised the overall creative quality online and celebrities have lost their pull, the micro-creator economy is the market's answer to the quest for relevance and authenticity. 

The fact that Stjepan's Hypefy can help you find, negotiate and set yourself up with the right ones to work with is a plus.

Unoptimised taste is a lovely thing to stand for, and very hard to spot in your own work. In Naples they say every cockroach is beautiful to its own mother. I have written a whole piece on why pride is a poor judge of your own work. The short version is that pride tells you nothing about whether the thing is good, and AI is a very efficient personal pride-boosting machine. Seen from the other end, AI slop is content with no mother at all, and the market is starting to notice.

Noël Bunting fixed the matter elegantly. Cathay Pacific's "Asia Just Got Closer", built on Artfair by Bria, made available an ethical AI art platform where artists opt in, keep control of their style and are paid every time it is used. A genuinely interesting counter-example to the AI equals homogenisation story, because it expanded participation without collapsing craft. The artists were still credited, and very much still paid. A decision somebody had to make deliberately.

Whether slightly off is the new beautiful, as Max Keane argued, or simply slightly off, you do need other humans, and their collective intuition well managed, to be the judges of it. Beware of the temptation of a closed-off, intimate conversation between you and ChatGPT alone.

Stand for humanity

Sara Denby reminded me I have power over the tool, and I loved her framing of agents as our ultimate brand activists. How we set up our agent, and what we ask it to find, will determine the reality that ends up in our baskets. We become the reality we choose to focus on, which is how your brain decides what you notice before you have noticed anything. Agents can now filter us towards legitimate D&I, sustainable and ecological choices, making life much harder for brands greenhushing, axing DEI initiatives, or running questionable supply chains.

And finally Mark Taylor, founder of Automated Analytics, moved me from unease to intrigue in twenty minutes by making a very controversial claim tangible. AI should and can add jobs. 

Until we work out a society where we can be happily unemployed and thriving, I add.

In a room where automation mostly implied headcount reduction, he made a brilliant growth argument: use AI to augment the team where correct attribution points, and grow the value it uncovers, instead of using it to shrink the headcount.

I always get a little uneasy when I chair AI panels where people promise optimistic job changes in lieu of job losses, as I think the absolute numbers will back up the latter. So it was useful to find Mark is not alone as the Guardian reported this week that the jobs apocalypse has not turned up yet. Mark said something related: use the tool well and it unlocks more jobs. 

Even so, Goldman Sachs put net US job losses attributable to AI at around 16,000 a month in April, and the Stanford HAI 2026 AI Index found employment for entry-level software developers down nearly 20% since 2024. I would ask the economists to sort it out between themselves, but I suspect it would require a panel, and I have chaired enough of those this summer. What I will not be doing is telling junior developers their training needs no rethinking. I do feel for anyone graduating into this.

And until we work out how to all thrive unemployed and enjoying extended holidays, family time or exploring our creative sides more, I will be writing from my own view point and treating your comments as my personal mini board, unpaid and unusually well informed.

I do think we can work it out. 

And when we do, it will not be because the machine figured it out. 

It will be because the collective did.

What's next?

We have covered what is worth standing for. 

Between saying and doing, though, as my good friend Sally would say, there is always the “ How” you need to figure out to perform well. That is where we will go next week.

If you want to say hi, connect with me on LinkedIn or follow me on Substak where I publish weekly. It’s free and I’d love to see you there!