From human interactions to fashion, furniture, travel, and political debate, social media now permeates many areas of life. The various platforms all follow the same business model: they market their users' attention to advertisers, and the longer users scroll, the more revenue the companies generate.
Hussam Habib is researching the strategies they use and the consequences this has. His research at the University of Iowa has turned him into a veritable algorithm detective. Through the APART-USA program of the Austrian Academy of Sciences, he moved to the University of Graz in October. In addition to the difficult political situation for researchers in the United States, there was a second reason for him to come to Austria: thanks to stricter EU regulations, scientists in this country can keep a closer eye on the tech giants.
“In the US, a large part of my work involved understanding how the platforms' algorithms work,” explains Habib. But since these are secret, he could only research this using so-called sock puppet accounts, i.e., fake profiles. In Europe, however, the Digital Services Act obliges social media companies to give scientists access to the algorithms. “The tech companies will probably fight back, but here in Europe we at least have a legal basis.”
Social media personalities
Habib plans to build on his research in the US with his work at the IDea_Lab at the University of Graz. There, he took an unusual approach: he gave the different social media algorithms a kind of “personality” and found that YouTube has an aggressive and extremely reactive personality that responds very strongly to likes and views.
The microblogging service X, formerly Twitter, prioritizes building a network. Who follows whom often counts more than the individual post. The personality of the community platform Reddit, on the other hand, acts as a curator. What matters to the algorithm is which communities and subreddits users join.
Nevertheless, one thing in common emerged:
“Anger and negativity work best on all social networks,” says Habib. “People engage longer with content that excites them emotionally. That's why it's more successful for content creators to generate controversial or negative posts.” Habib now wants to use effective access to the secrets of algorithms in the EU to find out: Why does certain content get a lot of reach?
Which stories are played out to many users by the algorithms and which are not? “If we fundamentally understand how certain content spreads on the platforms, we can respond in a targeted manner,” explains the researcher. Perhaps his research will help policymakers and maybe even the developers of these platforms to improve social media offerings and enable more diversity in the content displayed.
Micro-degree
Incidentally, students at the University of Graz can gain a comprehensive insight into the function and challenges of AI systems with the micro-degree “AI and Society,” which is open to all bachelor's students. Those who want to delve deeper into this subject can enroll in the master's program in Data Science.