AI Tracking in Advertising: How Worried Should Global Consumers Be?

The online advertising market is changing, at least in appearance, and many companies, either by intuition or by making a virtue of necessity, are changing their approach, making massive use of artificial intelligence (AI) to offer consumers advertisements targeted to their particular tastes. How worried should we be about this scenario?

There is a continuous race between regulators and authorities on the one hand and Big Tech on the other. As in the race between Achilles and the tortoise, in the well-known paradox, it always seems difficult to guess who will cross the finish line first.

Every time the regulators set new milestones to stem the effects of what tech companies have done, through the use of the personal data of billions of users, those companies change, mutate and evolve. But this should not be regarded as a defeat, far from it. This continuous chase fuels healthy competition between companies and pushes them to do better; to the ultimate benefit of users and businesses. At the same time, the authorities are pushed to improve, to invest in new talent and to hire new resources - the Italian Data Protection Authority (the Garante) has recently published a call to hire some AI experts, to give just one example.

In recent times, alongside the regulatory evolution that the GDPR introduced in 2018 – through a mix of increased sanctions, accountability and exemplary measures, with impacts both in Europe and in the wider world – companies whose business model is based on personalised advertising have had to constantly reinvent themselves.

As is well known, the last big blow came with the update of iOS, the operating system of Apple’s iPhones, which allowed users to easily refuse to be tracked in their online browsing, a move that led to losses that some social networks estimate to be around USD10 billion per year.

If, on the one hand, this choice has highlighted the possibility of consumers pre-selecting their online exposure when using smartphones online, but not only then, it has, on the other, meant that some actors now complain that they have lost the ability to create a complete profile of the consumer through their activity outside social media, and, above all, to verify whether the banner shown has then led to the purchase of a good, or a  subscription to a service. Obviously, for a company that advertises on social media, knowing that a certain banner led to the acquisition of a new customer will lead to further advertising on that platform.

As the online advertising scenario changes – although readers should bear in mind that the use of “cookie walls” (pop-ups requiring users to accept cookies or trackers in order to use a website) by Italian newspapers is still a matter pending before the Garante – companies are also changing their approach. Following the data minimisation principle of the GDPR, trying to do the most with the least amount of personal data possible, AI is becoming increasingly valuable and impressive. These platforms, be they social media site or video streaming services, already know a lot about their users, without the need to “chase” them on the other sites they go to.

Some companies have understood this for some time, others have discovered it more recently, making a virtue of necessity. This constitutes one of the risks, and at the same time one of the opportunities, of AI. The ability to analyse billions of pieces of data, coming from millions of users around the world, makes it possible to test, and quickly understand, whether the proposition of certain content is the best choice, which will raise the level of engagement or lower it, all by looking only at how users interact with that content. And, being short videos in the case of social media, the possible attempts, until the magic formula is found, will be almost numberless.

The news, then, is twofold. The first key insight is that, before too long, there will no longer be a need for cookies and other tracking technologies, instead it will be enough to look at what we do on the individual site or social media platform to offer us advertisements close to our tastes.

The second is that it is not yet clear how worrying this possibility should be.

Respecting the principle of data minimisation, as well as what is regulated regarding cookies by the EU E-Privacy directive, is only part of the equation. Respect for fundamental rights, the principle of fairness on the part of the data controller, should not be set aside just because less data is used.

Companies will still have to provide appropriate disclosure on how personal data is processed, not forgetting the additional transparency effort required by Article 22 of the GDPR on the processing of data by automated processes. And the better these systems become at guiding user behaviour, the more data controllers will have to demonstrate their goodwill by not hiding behind lengthy disclosures, instead being proactive in showing the risks and opportunities of these technologies. In short, they will have to demonstrate fairness.

This is a concept that may seem vague, as does accountability, but which is of paramount importance and will become increasingly so, given the ever more massive use of AS, as well as the paradigm shift that we may experience few years from now, in which our use of the internet ceases to be screen-based and become, instead, immersive.