Since the GDPR (EU General Data Protection Regulation) came into force on May 25, 2018, the internet search giant Google, which belongs to Alphabet, started to internally and intensively work not only to adapt itself to the GDPR, but also to create tools that mitigate the misuse of personal information from third parties. Such an effort was not just a privilege of Google, but also of companies that develop internet browsers such as Safari and Mozilla Firefox, which began to offer the option of blocking third-party cookies in their software.
As the GDPR and the Brazilian LGPD (Brazilian Data Privacy Act) define personal data as information relating to an identified or identifiable individual, cookies have become an additional reason for concern, since, when establishing one’s profile, it collects data that make it possible to identify a user through the junction. Cookies are thus considered as personal data and therefore subject to restrictions contained in both statutes. It is worth mentioning that cookies create a file that is stored in the user’s own internet browser.
However, Google seemed to have found a solution to replace cookies that serves the purpose of the companies without violating the personal data of third parties: the FLoC or Federated Learning of Cohorts.
The difference between FLoC and cookies is that FLoC establishes an analysis of people’s behavior on the web and anonymously classifies it into groups composed of other people with similar interests, instead of treating each person individually, as cookies do.
Cookies are predicted to be replaced by FLoC in 2022, but in addition to facing challenges for the reliable system performance, Google faced harsh criticism from some institutions such as the EFF (Electronic Frontier Foundation), which stated that, in practice, FLoC is just replacing of one tracking mode for another, creating new risks, since companies would only need to offer Google login services to link the information obtained by FLoC and thus identify the profile of those accessing it.
Strong competitors such as Microsoft and DuckDuckGo were also opposed to FLoC and even developed tools in their browsers to block it, on the grounds that FLoC is actually a maneuver by Google to protect its leadership in the control of advertisements.
Faced with all such adversities, Google has decided to abandon the FLoC development and implementation in 2022, declaring its death.
However, it is a mistake to think that Google would not explore some alternative measure. And so, Topics was born, an API (a set of functions and procedures that allow creating applications that access the resources or data of an operating system, application or other service) of topics! The difference with cookies can be seen in the image below:
Topics’ proposition is indeed similar to that of cookies, but with a focus on classifying websites, learning the user’s interests while they browse the internet. Thus, Google hopes to demonstrate that it is not interested in identifying individuals, but rather in their preferences when browsing. In this way, Topics would keep the data from the last three weeks of browsing history, with an initial limitation of 300 topics, with a plan to expand the topics to be defined by Google.
In this way, the company categorizes websites in pre-defined topics, being certain that, for websites not yet categorized, an algorithm in the browser may take over, providing an estimated topic based on the domain name.
When the user accesses a website that supports the Topics API for advertising purposes, the browser will share three topics that the user is interested in – one for each of the last three weeks – randomly selected from the top five in topics each week. The website may share this with its advertising partners to decide which ads to show to the user. Theoretically, this is a more private and secure method of deciding which ad to show.
According to Vinay Goel, Product Director of Google’s Privacy Sandbox, “With the Topics API, a user can only receive topics from a set of 300 options in total, making it much more difficult to identify individual users associated with a given topic. Furthermore, with five assigned topics per user per week – instead of one FLoC group – it will be difficult for two websites to use topics to identify a user as there would only be a 20% chance at most that two websites could receive the same topic for the user in a given week”.
Google also to clarifies that it will also provide users with much greater control and transparency than the current cookie standard, insofar as users will be able to review and remove topics from their lists, as well as disable the Topics API entirely as well.
Finally, how other competing companies, advertising companies and even advertisers will view Topics and the added value it will provide in protecting third-party personal data from being breached and the impact this will have on their sales will remain a question. For now, there is great skepticism, which normal is to some extent for any innovative technology in this market.