Algorithms based on users’ browsing history continue to bombard daily newsfeeds with recommendations. As creepy as they sometimes are in knowing certain things, the programs also have flaws. The models are, after all, a means of prediction.
But one programmer and entrepreneur decided to do something about the sensationalist news items or silly cat photos in his newsfeed. As a result, Henry Boldizsar created an app that gives users more choices on how they filter out the frivolous stuff and find what matters to them.
Boldizsar formerly worked for the meditation app Calm. In his role there, he helped build a rating system for their sessions. However, one night, the junk scrolling down his screen inspired an epiphany. “What if instead, we used ratings which reflect what people actually find important?” Boldizsar told Wired. “Like nuance and veracity?” he continued.
Boldizsar left Calm and began developing an app with a more effective algorithm. Dubbed Gem, the new application launched on Monday.
How it Works
Gem is similar to Apple News, in that users can choose story categories like “politics” or “culture.” However, two important features take Gem’s functionality a bit further.
First, users can evaluate what they read using adjectives instead of just clicking a simple “like” or “thumbs down.” Instead, they can choose superlatives like “nuanced” or “objective.” Boldizsar calls these labels “nutrition facts for the internet.”
In other words, the application cuts through the fluff and delivers what users personally value.
Second, the algorithm also gives users more control over the type of content they want to see. For example, they can filter articles by type such as “news,” “opinions,” or “tutorials.” While Gem still uses an algorithm to recommend a daily “gem” article that a reader will likely enjoy, it’s designed to avoid what Boldizsar calls, “echo chambers, where only your world view is repeated, and work in opaque ways, mostly out of your hands.”
A Growing Trend
Updating recommendation systems is becoming a trend for some of the major online platforms. Controversial content that pops up in peoples’ feeds often dictate these changes.
For instance, YouTube recently revamped its recommendation system because users complained about being prompted to watch videos with extremist views and, for lack of a better term, what amounts to child pornography (why that’s on YouTube in the first place is another issue).
Similarly, Facebook altered the way its newsfeed ranks stories to clamp down on the spread of misinformation. The social media site has received well-publicized criticism for allowing misleading reports on its platform.
Who might these tech giants turn to in order to keep their recommendation systems clean and relevant? Smaller apps like Gem and similar ones like NewsGuard (which rates news by credibility) have stepped up to fill this void. Furthermore, a former Spotify engineer is working on an app that recommends a few things for users to listen to or read each day.
Overall, prominent online platforms could learn a lot from these burgeoning upstarts. Moreover, with fake news becoming such an issue, the tech world needs to stay vigilant about improving its recommendation models.