One of the most original thinkers in advertising, Graham Wilkinson at Kinesso has delved into how generative AI and lookalike audiences are transforming the industry.

He discusses the pros and cons of employing these powerful new tools, how companies and consumers are likely to adapt in future and how he advises clients to respond to a rapidly changing landscape.


DN: How do you think advertising technology is likely to evolve from here?

GW: I’m having three or four conversations a day with clients about generative AI. I’ve never seen something have such a massive impact, it’s something that’s got every body going. People are trying to do things that are relatively ‘shiny’, but I think there are further steps to take before it becomes really useful.
Where Generative AI is concerned, factuality was an after-market add on. So, unless your content is grounded in the context of advertising performance and brands, people will use generative AI, but everything will have to be checked by humans. Right now, it’s not living up to the dream of faster, more personalised content. Once it can be grounded in the context of performance, we may see a bigger difference. For example, if we can say: ‘Write me something that elicits a response to buy a car, that could be monumental.

‍

DN: How do you feel about computers’ capacity for creativity?

GW: Technology is at a point where neural networks can work in the noisy, dynamic space of the real world. There are still limitations, but the bar is higher than in the past. AI can’t process data in the same way as humans do: essentially, AI uses statistics to guess what you are asking it for. That’s not how we understand creativity, it’s more about composition, whereas humans have an innate ability to apply abstract ideas to different fields. AI struggles with that.

But it is incredible how quickly generative AI has taken off. Chat GPT search will become a natural behaviour for the next generation, just as search engines have already become. I’m astounded by the uses that people have found for it.

DN: AI is increasingly blurring the distinction between fact and fiction. How should advertiser protect their clients?

GW: It comes back to the layers that you need to sit on top of large language models and generative AI. You need to ground it in fact and in brand guidelines. Big brands will use an off-the-shelf version of a large language model, then ground it in locations, audiences, in who buys their products, in brand safety, in facts. All these things have to fall into place before it is usable by brands at scale.
It’s possible that AI will act as a balance to information bubbles, because it can deal with the noise of the real world. Information funnels become narrow because they can’t flex to the real world, whereas AI sees a greater relation between a broader set of things. From our perspective, we’re working on a system that looks at the unintended consequences of audience choice, such as location, gender, race and political affiliation. If we embrace AI, we should put an equal effort into making the usage of data fairer. This is what we should be striving for, rather than using it to control everyone.

DN: Where do you seethe most fruitful application of Deep Learning in advertising?

GW: Deep learning is incredible at pattern recognition, that’s its key schtick. It can see who is buying a certain product, and who is not buying. The difficulty is in explaining how it works, especially considering the number of hidden layers that can exist in a neural network. This needs more work, so that people can understand how it does the things it does, then how to address this, not only in machines but in society.

DN: From your work looking at the impact of LLAs and generative AI on media, where do you see these areas developing in future?

GW: I see these becoming embedded into our everyday lives. For example, Microsoft Office Suite uses generative AI and it will become an assistant in most applications, I believe. If you want to write something, enhance something, start a thought process…That’s the utility it brings to everybody. I recently attended a conference where we had to spot whether an ad was created using orthodox technology or AI. We guessed right 53 per cent of the time, so it was the flip of a coin. What was interesting was that the most effective AI-generated ad, based on an effectiveness rating, was actually composed by a human designer, using AI components, so it was disqualified. The future, I believe, will be one where we try to rebalance the machine and human point of view. Businesses will move up and down this scale – there will still be a place for an entirely non-machine-based agency.

‍‍

DN: As a passionate cyclist and fan of Zwift [a technology which allows cyclists to compete with one another virtually], how far do you think that online gamification will change our lives?

GW: The fact that Zwift is now working with the UCI [Union Cycliste International] to create an e-sport Olympics shows what an amazing impact the company has made in a short time. It has taken something that human beings needed and met this need extremely well. For advertisers, this should be an ‘eyes light up’ moment, because through virtual worlds like the Metaverse, you can double your inventory – with real and virtual versions.
There’s no question about the validity of gaming, it’s becoming more and more embedded in our lives. With Zwift, it combines gaming with staying healthy, so I’m a big proponent. It means you can do things, like chat with friends on the other side of the world while you’re cycling, that you can’t do in real life. Brands should be part of this.

‍

Graham Wilkinson, Chief Innovation Officer, Kinesso, has been with IPG for more than six years, having previously worked as EVP, Product Innovation Kinesso, and SVP, Global Head of Product/SVP US Head of Product and Innovation for Reprise.  

Responsible for Data Science, Product Design, Research & Development and Custom Solutions, Graham also developed Kinesso’s “Shark Tank” program, Engine+, to spur creativity around innovation and integration amongst IPG employees with the internal competition.  

With 17 years’ experience working in product, innovation, digital, marketing, creative, and various business leadership roles across Europe, Asia Pacific, and North America, Graham’s career path as been led by the common goal of building stronger communities, and empowering people around the world.

‍