We’re thrilled to present an exclusive interview with Diarmuid Gill, Chief Technology Officer at Criteo. With over 25 years of experience in technology, Gill discusses the transformative impact of AI at Criteo, and how recent AI advancements have made the technology even more accessible to all — enhancing productivity and creativity. He describes how the company maintains a rigorous design process to guide ethical practices, while navigating the complex technological and regulatory environment.
Join us as we delve into the insights of one of the industry’s leading tech companies, exploring the future of AI in advertising and beyond.
IAA Benelux: Has the recent evolution of AI changed anything within your organization, and if so what?
Gill: I think the great leap forward in the industry caught even the experts by surprise. Access to advanced AI capabilities was suddenly in the hands of anyone with an internet connection, helping to bring the technology into the mainstream. Even people who previously had no experience with AI could benefit from its potential.
We have, however, seen a massive uptick in the number of questions we get on the topic, but we were well prepared. Criteo is an AI powered company with more than 18 years of experience in the field, so in many ways we were somewhat ahead of the curve. Our testing journey with generative AI began long before it became such a hot topic. We have 3,500 employees, and of course they want to use AI tools in their daily jobs. It’s our responsibility as a company to help them use AI responsibly, to educate them about risks and provide guidelines.
IAA Benelux: What do you see as the risks of AI for your organization and for the industry as a whole and what is your role / your organization’s role in mitigating those risks?
Gill: Generative AI has incredible creative capabilities, which presents both opportunity and risk. It’s an opportunity when that power is used for value generation but it’s a risk when it gets used for malintent or where biases are left in the data. We work hard to ensure that our products are built to operate ethically.
We firmly believe that regulatory bodies across the world need to be proactive and to take the topic very seriously. The most important thing for them to consider is to look holistically at what is in the best interests of the end users. Looking holistically means taking a broad view of the different aspects that would be affected by any regulations including choice, transparency, privacy as well as other economic and social aspects.
They should also consider the impacts on the entire ecosystem and not put things into place that might inadvertently preference very large companies with over-sized access to almost endless amounts of data. It would also be helpful if the relevant authorities in different parts of the world were to coordinate their responses so not to create a patchwork of different regulations that some actors could exploit.
IAA Benelux: How do you ensure responsibility over AI-systems used within your organization?
Gill: We have a very rigorous design process, where all approaches are reviewed openly across the organization. We then conduct AB testing on all updates, ensuring that they perform in production as we had intended. In parallel, in the very early days of our company, we implemented a privacy-by-design approach to ensure responsible use of data.
Finally, we also have a Product Ethics Committee, made up of employees from different teams across the company, where the goal is to guide and advise on how our products are built and used. This is a topic that our employee base feels very strongly about and as such it’s an important part of our culture.
IAA Benelux: What do you see as the opportunities for the use of AI?
Gill: From a productivity standpoint, AI is already helping to remove a lot of mundane work from our day-to-day work, making us more efficient and able to achieve more. I am in the group of people that feel in the medium term that AI will prove to be more of a huge boost to productivity rather than a threat to employment. It is very important though that people do not let themselves get left behind by these advances in technology.
Creatively, the opportunities will only be limited by our imagination. I foresee AI unlocking new types of media, helping to make ads more diverse and more engaging for users. Shoppable video is a great example, using AI algorithms to personalize content shown and track objects within the video, enabling direct links to product pages when viewers click on specific items. This could easily transform the lower funnel capabilities of video-led channels.
IAA Benelux: Do you see the industry using AI for the collective good, and if so how?
Gill: People have had a lot of fun playing with the power of generative AI and new applications are popping up all over the place. The amount of creativity has been amazing to behold. I think we are still caught somewhat in the hype bubble, so we won’t really know what the long-term changes will bring until we exit this first phase of the generative AI revolution.
I can see many practical applications that will benefit society. For example, in health, a practitioner can use these tools to spot things that might escape the human eye. These technologies will be an aid for such people but will not replace them. I think we are still a long way from being able to depend on such technologies, but for already skilled professionals the potential of these tools is incredible. Imagine a doctor being able to review multiple times more patients than they can currently. Speedy diagnosis is very often the most important factor in achieving a successful outcome for a patient.
IAA Benelux: How are the regulations around AI, such as the EU AI Act impacting your business in the EU and abroad? Do you see these regulations as opportunities or threats?
Gill: At Criteo we have always been very keen to ensure that AI is used for good and that we don’t accidentally create systems that inadvertently discriminate because they were trained on biased data. We encourage any regulations that will help our industry counter this. Equally, it’s important that regulations protect against unfair playing fields, where large tech players who sit on vast amounts of first-party data could take advantage.
As the world’s first ever AI legislation, we’re very hopeful that the EU AI Act will deliver much needed standards and regulation around the highest risk AI systems while protecting our fundamental values. As the first EU private AI lab based in Paris, Criteo stresses the importance of enabling AI innovation in Europe through open-source AI models and systems.
We need to ensure that it remains possible to work with such freely accessible models. Otherwise, generative AI will become the private property of large corporations that have the financial resources to train such AI models on their own.
However, there are certainly still some unresolved challenges for companies that want to use generative AI. For example, how the protection of intellectual property will be enforced by the courts, and which forms of use by generative AI are to be permitted. In the coming months, we will gain a clearer perspective on how the contribution of different parties should be assessed, at least from the US perspective.