Q&A: Shruthi Chindalur, executive managing director EMEA, Criteo

Criteo's Shruthi Chindalur discusses the future of AI technologies and the ethical challenges they present.

Shruthi Chindalur, executive managing director EMEA, Criteo

Over the past six months, consumer habits have evolved at great pace. According to personalised retargeting firm Criteo, 49% of UK shoppers have discovered a new form of online shopping during lockdown.

This dramatic increase in online consumer activity has precipitated a shift in companies’ marketing operations. Marketers are increasingly turning to Artificial Intelligence (AI) strategies, and chief among the technologies to have gained momentum in the sector are machine learning and big data.

It is, therefore, essential that marketers have a robust understanding of AI technologies and how to apply them.

Shruthi Chindalur, executive MD EMEA at Criteo, speaks to Marketing Gazette about these technologies and the ethical issues surrounding them, as well as the impact of COVID-19, and Google Chrome’s third-party cookie removal.

What impact do you believe machine learning and big data will have on the world of marketing?

We’re living in an era where people have never spent more time online. They have information on tap, and shifting consumer behaviours in response to the COVID pandemic have only thrown this more sharply into focus. What has this got to do with marketing? Well, at its most basic level, marketing is a conversation between someone who is selling something and someone who is looking to buy something. So, as consumer shopping habits are changing and marketers have more data than ever to play with, machine learning and AI have a pivotal role to play in making sense of all this information.

For example, AI systems help marketers understand new and prospective customers in ways that go beyond crude socio-demographic audience data so that messages can be tailored to an individual’s needs, desires or concerns. But this is only the beginning. AI has the power to help marketers bridge the gap between the online and physical worlds by using time or location sensitive data to better understand their products, predict which ones will sell best, who to, at which price point and through which channel. Because AI is relatively easy and cheap to implement, we’re beginning to see the democratisation of AI and machine learning in the world of marketing. Where previously it was an innovation reserved for big business and large national retailers, smaller and more bespoke retailers can also now reap the rewards of reaching their community online and use AI to gain a competitive edge.

To what extent do you think this impact will be affected by the coronavirus pandemic?

The coronavirus pandemic has hugely accelerated the trends and shifts we’ve seen taking place across e-commerce. For instance, our research shows that half (49%) of UK shoppers say they have discovered a new form of online shopping during lockdown, with 81% intending to continue this behaviour in the future and once the pandemic is behind us.

This is a hugely important shift in behaviour that’s seeing people purchase goods online that they used to buy in physical stores – and it’s here to stay. It shows that the power of digital cannot be underestimated and the role of technologies like machine learning and AI will only become more important as retailers large and small must ensure they have the best in class solutions for online right now.   

Some people believe machine learning should be paired with other AI systems that encode knowledge and are capable of reasoning – classical or symbolic AI systems, for example. What’s your view on this?

This remains a large area of debate within the AI research community – particularly in terms of interpretability – and is still largely in its infancy when it comes to applying it to the world of marketing to deliver impact. For instance, large AI models require an enormous amount of training data from which they’re able to learn and refine analysis based on the experimental data provided.

However, merging existing knowledge and learned parameters, as well as mixing AI-driven tasks with those traditionally done by humans, is very difficult and will likely always require an element of human supervision and compassion to get right.

Google Chrome has begun to phase out support for third-party cookies. What impact do you think this decision will have on digital businesses and technological innovation?

This question fundamentally comes down to commerce as a whole and the impact it will have on the economy. Google’s move has come as the world of digital advertising has seen a significant shift in how ads are bought and sold, with a greater emphasis on user privacy including calls for more transparency, choice and control over how users’ data is used.

This is not new, and clearly digital businesses need to evolve to meet these demands. Ultimately, it comes down to empowering users by giving them control. How? Our proposal ‘project Sparrow’, outlines a privacy by design protocol that is totally secure and based on user choice to help resolve issues around data bias, anonymity and security. It aims to enhance Chrome’s proposal by providing more control and transparency while maintaining privacy guarantees for users. We really hope this proposal is given a serious consideration and acceptance.

There’s a widespread fear that machines will one day be able to learn too much for themselves. The use of AI also raises many ethical issues for society. What are your thoughts on this, and how do you believe public confidence in AI technologies can be maintained?

In the short-term, building trust and public confidence in AI technologies has to come down to the way systems are developed in the first place. Oftentimes bias, discrimination or lack of fairness in data sets or AI systems is a result of the way these are built. In the long-term, widespread acceptance of AI technologies as a part of daily life will come down to how transparent and open businesses are in the data that is collected and how it is used. For instance, the privacy issues we touched upon earlier will only become more important as consumers are empowered to choose how their data is used on their terms.

What ethical limitations, if any, do you think there should be on the use of AI?

Getting the process right in terms of determining where, when and how ethical limitations should be placed on the use of AI is fundamental. Nailing the process means having the right conversations with the appropriate bodies. For instance, Criteo’s AI Lab regularly engages in conversations with the European Commission, the CNIL and French and European Members of Parliament to ensure that privacy and data protection issues are always at the forefront of any new research, project or innovation.

Ultimately, we believe that the strength of any new regulation should be proportional to the impact on people’s lives and fundamental rights. It’s a difficult balance to strike but one that is incredibly important to make sure any new regulation of AI does not hamper innovation in the field and widen the technological gap between Europe and its competitors.

  • Criteo is a global technology company that works with e-commerce companies to assist marketers with trusted and impactful advertising using machine learning and big data.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.