Google has shared its plans to eventually move away from using cookies for building user profiles for marketing. Instead, the company wants to use artificial intelligence to show users relevant ads.
To this end, the company has built a new AI called Federated Learning of Cohorts (FLoC). Google claims that the AI is almost as capable of showing relevant ads as cookies are, but that privacy is better protected.
AI runs on user’s computer
With FLoC, calculating the user’s interests is left to the user’s computer. Based on the website the user visits, the browser compiles a general profile of previously visited websites, which Google calls a ‘cohort’. Websites can request this profile and deliver advertisements based on this profile.
Thousands of the same profiles
Such profiles are not unique. Eventually, thousands of people will get the same profile, making it impossible for a website to distinguish individual users. Google also does not want the cohort to be marked with more information than a short identification number, such as 43A7. The company promises that the ads will be up to 95 percent as relevant as when tracking cookies are used.
More privacy-friendly than cookies
In combination with other data, such as an IP address, websites might be able to draw up a broader profile of a user, but this solution promises to be a lot more privacy-friendly than the current solution using cookies. With cookies, websites get a fairly broad overview of your browsing history to base their decisions on what ads are served on.
Several browsers already block these tracking cookies, including Firefox and Safari. Google’s own Chrome browser still collects tracking cookies, but the company plans to phase that out in the next two years. Since the company is largely dependent on online advertising costs for its revenue, it must first have an alternative for the cookies. FLoC could possibly become that alternative.