AI hasn’t pressed delete on creative and responsible advertising  

Working as a creative leader throughout her career for the likes of BBDO, Karmarama, Cheil, Meta and now Dentsu, Ryan has guided teams through multiple waves of cultural and technological change. And, lo and behold, the core ethical responsibilities remain the same: “don’t create harm and be honest.” 

In its early phase, AI experimentation was not commercially viable, but creatively it hinted at what was coming. “It was exactly the same when Photoshop was first introduced,” Ryan says. “The time taken to reach a result shortened, which meant experimentation exploded.” 

Fear has always existed in the early days of technological revolutions. But what will help the industry move forward, Ryan began our chat by saying, is not “throwing out the guardrails” but maintaining commercial responsibility - “just as agencies always have” - while embracing AI’s creative potential. 

Because AI systems optimise so fluently for performance metrics, they can create a sense of authority or objectivity - which is why human judgement remains critical. AI can sometimes seem like it has a value system we can rely on, Ryan admits. While the tech may often work through a performance metric and not a human metric, it can be easy to side with its decisions. This is where Ryan feels creative leaders really need to lean in.

“You can’t abdicate responsibility.
Humans still have to be involved."

"There’s a temptation to think, ‘The machine came up with it, it looks sound, so I don’t need to intervene’ - and that’s precisely when human involvement matters most.”

Ryan believes the industry has a role and responsibility to be transparent and avoid misleading and over-claiming. “Responsibility doesn’t disappear just because AI comes into the process.”  

In every major innovation - from early transport to modern automation - safety has had to be built in. Understanding where safeguards are needed still matters, Ryan says. “Don’t assume that platform-level safeguards replace human responsibility, even when guardrails exist. Judgement still sits with the people using the system.”

Transparency also plays a key role in how brands connect with consumers. The creative veteran calls for agencies to openly acknowledge - rather than quietly conceal - their use of AI agents. “Stepping ahead of it and showing it's part of your workflow demonstrates strength, not weakness. But at the moment, I have heard from many, ‘Oh, but isn’t it cheating?’” 

Ryan builds her own task-specific agents, designing prompt-driven systems and workflows rather than training models from scratch.  “I build agents all the time, and you have a lot of control. You see the output, you realise it's not what you wanted, and you go back in and rewrite the prompts or adjust the temperature. You are the architect of that experimentation and that process - you're not divorced from it.”

Whether agents are custom-built or pre-trained, Ryan is clear: accountability never shifts to the technology. Users must still govern what the system interrogates, how it behaves and whether it adds value. “You still have to ask: ‘Does this look right? And is this good for society if I put this ad out into the world?’” 

With regard to Dentsu Creative’s use of AI, the agency’s partnerships with Adobe, Google and Microsoft’s Copilot focus on transparency - and on clearly understanding what each tool does and does not do. 

“Adobe Express is a commercially clean environment because it is trained on licensed, owned and public-domain content, which gives greater commercial certainty. You know you're not using somebody else’s work or failing to credit the right artist. With less transparent training models, such as Midjourney, the aesthetics might be very pleasing but you're not always sure where the commercial licence starts and ends. That’s a very important part of Dentsu’s process.”

Ryan stresses the importance of embedding AI use into the workflow rather than allowing it to happen in isolation. “That’s where you can get into trouble,” she says. “Our creatives use tools like Firefly collaboratively - with clients and teams - not in silos. We look at the workflow end-to-end, with check-ins, sign-offs and the same legal and brand reviews we’ve always had.” 

AI can replicate cultural biases because it is trained on human behavior, which already contains bias. To prevent this, Ryan emphasises that leaders must stay aware and not accept outputs at face value. While early visual artefacts are becoming rarer, more subtle risks remain, particularly around reinforcing bias, stereotypes and narrative shortcuts.

“You are in control. If your prompt isn’t specific there’s a risk that the machine which is already looking for patterns will replicate human stereotypes, because that’s exactly what stereotypes do - they shorten the cognitive load to create something that most people shortcut to. Being aware of this and designing to avoid it is largely in the user’s control – but only if people take that responsibility seriously.”

Hyper-personalisation is another area where AI-powered advertising risks crossing ethical lines. Ryan is particularly excited by the power AI has given to personalisation but she also recognises that it has potential brand-damaging effects. “I’m a big fan of relevance,” Ryan adds. “You can suddenly make brands super relevant to different audiences in a meaningful way. But when does that become exploitation?” 

Advertising crosses that line when data is used in ways people would not reasonably expect - through lack of permission, opaque consent, or manipulation designed purely for financial gain like exploiting someone’s race, gender, marital status or search history.  

“It becomes incredibly important to protect data, and to make sure that when you're creating relevance between a brand and a customer, you're doing it in a way that adds value rather than stress, guilt or emotional distress.” For Ryan, being human about how you use the tool, and being protective of the end users or viewers, is incredibly important. 

While onus can be placed on values and social purpose, advertising, of course, exists to creatively capture imagination to sell products and services. Ryan believes creatives have drifted from that commercial outcome - but that AI can help reconnect the two. 

“Creating hyper-relevant content that presents a brand in a way that is meaningful to somebody - not just shoving it down their throat - is far better than the idea of something chasing you around the internet based on demographics or psychographics and trying to sell you a product you have no interest in. That’s not good for the advertising industry. 

“If we can turn that on its head and make sure that when a brand appears in someone’s space, it does so in a meaningful, honest and transparent way - and genuinely fits that person - then they’re far more likely to buy. That’s a good outcome.” 

Ryan’s enthusiasm for AI stems from its ability to accelerate experimentation — a core part of creativity that allows ideas to be tested, pushed and reshaped quickly. 

While many creatives use AI to generate a single, often unoriginal solution, Ryan encourages her team to “go way out.” They experiment with temperature - from precise outputs to letting the AI “go wild” and form unexpected connections. 

"It’s a useful way to think about it. Are you drilling down to one answer? That’s not great for creativity. Don’t make the mistake of treating probabilistic systems like search engines rather than tools for exploration."

AI also enables creative “mash-ups” - combining unrelated genres or ideas to spark invention. “That’s where new ideas come from. AI lets you do it affordably and efficiently. The Pope and Balenciaga moment worked because it surprised people emotionally — two worlds colliding in an unexpected way.” 

Ryan advises agency leaders that above all, they must have the mindset that AI can be used to create distinctive, authored and culturally intelligent work - “rather than just commoditising and creating a whole load of shit,” she jokes. “I think AI can actually provide the opposite.” 

“Capabilities across models are converging rapidly in many areas, even though each system still has distinct strengths and weaknesses. It’s valuable to experiment across models to understand their strengths, limitations and appropriate use cases."

“Less emphasis needs to be placed on knowing every tool inside out. In a lot of the work I do, the biggest barrier isn’t technical understanding - it’s mindset,” Ryan says.

“No matter what the platform is, the question should always be: how do we use this responsibly, creatively, and with intent?” 

For Ryan, that intent is clear. AI is not here to replace creative judgment or ethical responsibility - it is here to sharpen them. Used well, it can empower creatives to experiment more freely, connect more meaningfully, and build work that serves both brands and people. The guardrails have not changed - but the opportunity has expanded. 

“What excites me most about this moment isn't the technology itself - it's the strategic questions it forces us to confront. As these tools become more sophisticated, I believe we'll see new models emerge for how creative and strategic thinking gets packaged and delivered. The agencies that thrive will be those willing to fundamentally reimagine their value proposition, not just bolt AI onto existing structures. For me personally, that's the puzzle I'm most energized to solve."