The web runs on algorithms. Your search results, product recommendations, and the news you read are all customized to your interests. They are designed to increase the time you spend in front of a screen, build addiction to sites and services, and ultimately maximize the number of times you click on advertisements.
Without discounting the utility that this personalization can provide, it’s important to consider the cost: detailed portfolios of data about you are sitting on a server somewhere, waiting to be used to determine the optimum order of your social media feeds. Even if you trust that the parties collecting that data will use it responsibly, it has to live somewhere and has to be transmitted there, which makes it a juicy target for bad actors who may not act so responsibly.
At Mozilla we think the web deserves better, and we believe that we are uniquely positioned to offer you the best of both words:
A number of ongoing Firefox projects attempting to provide these benefits with Mozilla’s sensibilities:
- The Pocket team is experimenting with a privacy-conscious approach to sponsored content.
- Intern Florian Hartmann is investigating use of a machine learning technique that does personalization locally, only sharing models with a central server, rather than the data used to produce those models.
- The Firefox Pioneer program allows users to opt-in to sharing their data with us to help us develop these products.
Advance offers you a new type of forward button, making real-time content recommendations from elsewhere on the web.
Advance offers contextual recommendations based on the pages you are currently visiting.
Over time, these recommendations will become personalized to your interests, learning from your interactions with the experiment and the broader internet. In addition to recipes, try it out with book reviews, blog posts, and news stories. We think Advance will help you find your next favorite thing.
Over time, recommendations become personalized to your interests.
We’re launching Advance in collaboration with Laserlike, a machine learning startup that has built a web scale content search, discovery and personalization platform. They’re our trusted partners on this project, and we’re so grateful for their help in advancing our mission.
We’re trying to prove that we can use these technologies in the right way, and refuse to sacrifice user control to do so. The experiment is opt-in, and at any time you may pause its data collection. You are able to view the data that Advance has collected about you, and may request the deletion of that data at any time.