AI tools have been going through research and development for decades now, but the recent rapid commercialization of tools such as ChatGPT or image-generator Midjourney have brought about a level of public awareness for AI not seen before. While these tools may seem, and to some extent are, revolutionary we are clearly only just skimming the surface of the capabilities of AI.

In this interview, David Roth, CEO of The Store EMEA and Asia, the global retail practice of WPP and Chairman of BrandZ and BAV Group, discusses the impact of AI tools on the branding and consumer industries, the need for legislators to catch up with current AI technologies and the fact that we are at the beginning of what is likely to be a hugely developmental period for the world as a whole.

Q. To what extent would you say the branding and consumer facing industry has been impacted by the commercial availability of AI models such as ChatGPT?

A. It is very contextual, because I think all industries, including the advertising industry, have been caught out. Not by AI itself as it has been around in some form or another for 20 or so years, but because of how quickly it’s become consumer-focused. ChatGPT is a good example of AI moving from a research environment to a commercial environment. On top of this, the other thing that’s been quite extraordinary is the speed of adoption, both from a consumer perspective, and also from a commercial perspective of how quickly people are starting to do some really interesting things that have commercial, and obviously from my particular discipline, marketing applications so you can now go from having a thought to implementing it in five minutes.

Q. With the speed of change toward commercial availability, how soon will it be before we see the true impact of the technology?

A. I suppose because of the speed, we’re forgetting that the adoption of AI applications is still in the nascent phase. The impact this is going to have is going to be massive and I think we are yet to scope out exactly what that impact will be. There are some existing applications, but for the most part we are seeing glimpses of potential applications and some of those are very distant right now. It’s an amazing proposition and we are going to be capable of creating things that have never been able to be created before, but you also realize that we are functionally still at day one. The direction is clear, but the speed of travel and what will appear along the way are still very unclear. That’s both very liberating and very scary. If you look at this from a political dimension, I’m not certain there’s been an adoption of technology that has exercised the political environment as much as this has in such a short space of time. It usually takes politicians and governments 10-15 years to get to grips with new technology and start to understand what the legislative framework around it is going to be. But you can’t do that with this, it’s pretty obviously an unstoppable trend.

Q. What are the benefits and drawbacks of using AI tools to create marketing content across mediums and how does this compare to solely human generated content?

A. You can split this into a number of areas. The first is clearly speed. AI tools will speed up all sorts of processes, not completing them, but taking you from, for example, zero to 50% much quicker than a human currently can. And in a few years, when people have optimized the desired outputs, it could well be something closer to zero to 80%. Of course, with speed also comes productivity. But with productivity there is also a potential reduction in employment, and that’s a huge issue that needs to be addressed.

Another impact is the ability to use the technology across various different parts of the marketing mix and join them together in a coherent and cohesive way. This will have an impact on creativity and craft skills, fast-tracking new marketing environments and creating deeper relationships with consumers. For example, the digital practice of using 2D objects to create 3D objects to simulate different environments has been unbelievably prohibitive. But this type of technology totally simplifies this process. That is just one example, and the more of them there are, the quicker the move to a more affordable and immersive type of marketing platform.

Q. To what extent are there real risks to transferring responsibility of content generation to AI tools?

A. We have so much universal visibility of products through things like Amazon, Alibaba and JD.com. But the brands attached to them denote a certainty of expectation and have great importance because of the trust they hold. People often purport that brands have become less important because of the information consumers have at their fingertips these days, but actually the reverse has happened. Brands are becoming more important because of the sheer amount of choice available and consumers do need and value curation of that choice, and they trust brands to do it.

If you add AI into the mix, the questions then become: What is the trust relationship we’re going to have with artificial intelligence? Who are we trusting? Are we just trusting it because it’s new? I think perhaps we will end up with delineations of AI tools, maybe groups like the EU, trusted existing brands or even a completely new brand could certify trustworthiness via some sort of certification.

It will also be important to make sure that the data going into the AI tools is as bias-free as possible. People have inbuilt bias, socialized or otherwise, and transferring this into the function of AI tools even unintentionally could have serious long-term consequences. Ensuring that computer vision algorithms, for example, are universal has been a big topic of debate, especially as you start moving the models from country to country.

Q. How are you using generative AI models in your day-to-day work, whether it be companies you work with or you personally?

A. Details are hard to say, due to confidentiality, but what I would say is that everybody is experimenting, either out of interest or fear thanks to some board director grilling them on what they are doing about AI. In terms of broad categorizations, there is the use of AI for improving internal processes and helping to understand and use corporate knowledge.

Then there is the external, which is around the consumer proposition. So there are different types of consumer propositions that start to come out of these technologies which enable brands to have a deeper, more meaningful and more trustworthy relationship with their consumers. This might see the use of AI to assist with analyzing customer feedback to improve product development and customer-facing processes.

Q. Who are the main players in terms of generative AI technologies, both in China and internationally?

A. As with any disruptive technology, it provides an opportunity for new players to suddenly enter into the mix in an environment that was originally shut out to them by existing brands with good positioning and plenty of resources. But something like this has the capacity to be phenomenally disruptive. If the current players of today, like Microsoft, Google or Baidu, don’t cement their position in the space, they will find it significantly weakened and we’re likely to see a whole new set of businesses in a few years’ time.

The question for me is whether there is going to be a baseline technology stack that everybody dips into in order to run their particular applications or will it be the main players, because this is a very expensive game, out there using their own base level in order to create things that are differentiated and distinctive?

If you look at it from a global perspective, this is something that will have to be global. It isn’t something that can just be developed by the UK, US or China, it’s going to have an impact on every single country around the world. It may well actually be something of a leapfrog technology, like with China going from no phones straight to mobile phones, that helps springboard development in countries in the global south. For the global north, we’ll see rapid development and for China in particular, it used to develop 10 times faster than anywhere else, but it could almost be double that at this point.

Q. One of the reasons China is so advanced in their fintech development is that they leapfrogged credit cards, while countries like the US and UK are slowly weaning themselves off them. Do you foresee any sticking points like this for AI development?

A. The optimum use of AI technology is when you have complete transparency in what’s being fed into the various different models. So I think there is a big question about the freedom of data pools in various different countries and how easily AI can access and ingest data. It will also affect which AI engines Country A or Company A choose to use, depending on the transparency and universality of data. Again, this is probably somewhere that needs politicians and legislators to step in.

Q. Looking forward, what other opportunities do you see for AI?

A. My hope is that this type of technology enables the world to become more equal. It is hard to envisage how that might work just yet, but we have a massive opportunity to level out the disparity between even the very poor and the middle class in countries around the world that have so far been unable to take advantage of the various technologies invented over the last few decades for whatever reason. I think it can be a fantastic contribution to humankind. ChatGPT has done an amazing job in terms of commercialization and there are a massive amount of other applications being developed on various data sets, but we have to be careful of where this all goes and really get ahead of developments because they will come thick and fast and probably exponentially, and we are already quite far behind in terms of understanding and especially in terms of legislation.

Interview by Patrick Body