In the sphere of digital media, AI is no longer in the experimentation phase; it’s becoming key infrastructure. And that brings a tension with it – real optimism about what AI can do, set against equally real concerns around job displacement, ethics and the environment.
When speakers from agencies, platforms and publishers convened in London for the AI Growth Summit, they did so because of a growing sense of urgency around AI adoption in advertising.
That urgency ran through the whole day, with sessions focused on a few core themes: trust, transparency and accountability. And two questions in particular kept surfacing: human creativity versus automation, and the shift from attention to intent.
Is the human just in the loop – or completing it?
In her philosophical keynote, Ogilvy’s Amelia Torode spoke of 2026 as a time of overwhelming information but limited wisdom. Her point was that while AI hands us instant answers, all that speed can quietly rob us of the learning that comes from working things out ourselves. She likened AI tools to taking a helicopter to the mountaintop: you get the view, but you miss the journey that gave solving the problem its value. This is exactly why human judgement becomes more valuable in the AI era, not least because we need it to make sense of this period of accelerated and profound change.
Arguing that travel advertising is one of AI’s biggest opportunities, Tripadvisor’s Rachel Vardon also spoke to the essential role of human input. With the company well-placed in terms of consumer trust and data, the possibilities that AI offers are extensive: conversational travel assistants, AI-generated travel videos, and enhanced content discovery. But she also stressed that direct human contributions – reviews and lived experiences – remain an essential part of the overall advertising puzzle.
Most AI today is predictive pattern recognition, stated Tom Goodwin. Brilliant for coding, research, planning and market analysis. But lean too hard on AI for production efficiency and you just end up automating mediocrity at scale. Once again, the human touch is what brings the emotional storytelling and brand-building that actually sticks.
That human role comes with responsibility, too. Zehra Chatoo of Code For Good Now used her session to interrogate AI’s blind spots, warning that models simply mirror the data and the people behind them. Her research identified what she called an “AI judgement penalty”: women are adopting AI at roughly 20% lower rates than men, held back by concerns around trust, ethics and fear of being judged, and identical work is rated more harshly when people are told AI helped produce it. Her conclusion was blunt: the stakes are too high for AI to be trained on narrow perspectives and narrow participation.
Chalice AI’s Freddie Turner brought the most practical note of the day, arguing that AI should work for the advertiser, not the platform. She drew a pointed parallel with programmatic, which arrived a decade ago promising transparency and efficiency but delivered black boxes, arbitrage and a lot of broken trust. With “emperor’s new clothes” solutions everywhere, her advice was to interrogate anything being sold as new AI with a handful of blunt questions: can you explain how it actually works and which signals it weighs? Whose data is it using, and is your competitive advantage being quietly commoditised across other advertisers? Who owns the resulting model, what is it optimising towards (your bottom line, or a platform metric like CTR?), and can you take it with you if you leave? Trust, as she put it, is hard to build and very easy to break.
From search to conversation – and from attention to intent
Conversational discovery is replacing keyword search, and AI assistants are becoming the gatekeepers to consumer decisions. In fact, nearly 60% of consumers are ready for AI agents to replace traditional websites for tasks like shopping and travel planning. So said Microsoft AI’s Tracey Pilon, whose central argument was that brand visibility therefore increasingly depends on being understood and trusted by AI systems.
Captify’s Sam Coates and Emma Welch also explained that search marketing is evolving from reactive to predictive, with LLMs becoming the new search ‘front door.’ What we share or express to a chatbot might vary from what we discuss with one another. Their proof point was striking: consumers recommended a brand by an AI model were nearly eight percentage points more likely to consider it once they reached the open web. As they put it, no one lies to their search bar, which makes these early signals unusually honest. This allows for earlier insight and an alternative perspective into consumer interest, and advertising is becoming more intent-driven as a result.
Nano Interactive brought all of this down to earth with a showcase of NanoQ, their agentic media planner. The tool is centred on the core idea that consumer intent is communicated across multiple behavioural signals, before a purchase decision is made. Analysing sentiment and contextual relevance in line with a brand brief, it builds audience segments and exports campaigns directly into ad platforms. The AI works as an intelligence layer, with human oversight, while remaining both explainable and transparent.
The era of AI as an operating system
The urgency has hard numbers behind it. The IAB’s own research, referenced on the day, positions UK digital advertising as a rare growth engine and suggests AI-powered ads could unlock an additional £12 billion for UK businesses in the coming years, with AI-driven formats potentially accounting for close to a third of digital ad spend by 2030.
Bloomberg’s Parmy Olson offered a sober counterpoint from inside the AI industry itself. She pointed to a growing public backlash – graduates booing AI executives off commencement stages, teenagers anxious about the technology’s environmental toll – and cautioned against dismissing that discontent. On advertising specifically, she noted the big players are only beginning to experiment with monetisation, from banner ads beneath answers to sponsored replies, and warned that dropping an ad into the answer itself would be “playing with fire” with users already wary of how their data is used. With the economics of these models still wildly out of balance, she expects far more ad experimentation over the next couple of years.
AI is becoming embedded across every layer of advertising, and the message throughout the day was this: the future belongs to organisations that a) consciously combine automation and machine learning with human creativity, and b) understand how AI can allow advertisers much earlier access into the consumer journey.
And what’s the objective that unites them? To enable operational efficacy and speed, and in parallel maintain transparency and trust.
One last thought lingered after the sessions wrapped: the headline shift here might be less about automation and more about consumers handing their decisions over to AI. Because AI has stopped being a future trend to keep half an eye on. It’s quietly becoming the operating system that modern marketing runs on.