“There are a couple of improvements in Topics,” says Hamed Haddadi, chief scientist at Brave, a privacy-focused browser and search engine. He says that under FLoC, people could have been grouped into more than 30,000 different categories, which would allow advertisers to gain specific knowledge of their interests. This information could then be combined with other data to build up an incredibly detailed picture of each and every one of us.
This is less likely in Topics, as there are around 350 interest categories that can be assigned to people. Although this number is likely to increase—Google’s technical description says its eventual goal will be to source these topics from a third party, and there could be a “few thousand topics.” Haddadi also says adding a sixth random topic into people’s interests makes the system a little more privacy-conscious.
Another potential difference between FLoC and Topics is that Google claims the latter will attempt to avoid assigning “sensitive categories” to people—such as allowing individuals to be shown ads based on their race or gender. FLoC was criticized for potentially being able to generate or infer sensitive attributes through people’s behavior and interests. Google says people will be given more control over the interest areas that are assigned to them and can change settings, block topics, and opt out in Chrome. But, realistically, it’s unlikely many people will change Chrome’s settings in this way.
What’s more, the risk of websites working out someone’s sensitive personal traits isn’t completely eradicated by Topics. “It is still possible that websites calling the API may combine or correlate topics with other signals to infer sensitive information, outside of intended use,” Google’s description of Topics says. Over time it would be possible for a site to “develop a list of topics that are relevant to that user,” and this may reveal sensitive information. There are other privacy and security issues Google says it needs to fix. Google plans to test Topics in Chrome in the coming months, and the system could change based on feedback.
Then there’s the competition issue. The smaller number of interests assigned to people could potentially hand yet more power to Google in an online advertising industry it already dominates. Paul Bannister, cofounder of the ad management firm CafeMedia, says that Topics seems to be a step forward for people’s privacy, but a potential step back for advertising firms. The 350 current interests included in Topics are broad, Bannister says, and this means it’s less likely to be useful for advertisers who are trying to target individuals with products that they’re more likely to buy. “Those topics are fixed, so it’s harder to find unique segments that are really interesting to your marketing campaign,” he says.
“As it stands, Topics seems to be only a solution for the Chrome browser. It is neither cross-browser nor cross-platform,” says Phil Duffield, UK vice president at the Trade Desk, a tech and software company. The company built its own cookie-replacement rival that is based on identifiers linked to the email address people use to sign in to websites. “As with any complex technical challenge, there is no silver bullet, but we do believe in the importance of future solutions being interoperable and easily used by all players across the industry,” Duffield says.
.
