European Union regulators have taken action against major social media platforms, including Meta‘s Facebook, TikTok, and X (formerly known as Twitter), in response to concerns surrounding misinformation and “illegal” content related to the Israel-Hamas conflict. The EU has given these platforms a tight deadline of 24 hours to address the issue. However, navigating the intricacies of content on these platforms, especially TikTok’s algorithm-driven endless scroll, is a complex challenge.

The situation was exacerbated when, following recent events in Israel, X became inundated with false information about the war. Verified users, recognizable by their blue checkmarks, shared videos from past conflicts, footage from unrelated parts of the world, and even gaming clips disguised as authentic coverage from Gaza. Such content was amplified by the platform’s algorithms. European regulators, led by EU Commissioner Thierry Breton, quickly demanded action, accusing these platforms of violating the Digital Services Act.

Breton subsequently directed similar letters to Meta and TikTok, highlighting TikTok’s unique responsibility to protect children and teenagers from violent, extremist, and potentially life-threatening content. TikTok acknowledged the presence of such content and claimed to have bolstered its efforts to combat violence, hate speech, and misinformation. The company reported employing a team of 40,000 “safety professionals” to review content and collaborating with the International Fact Checking Network to counter false claims.

Misinformation is a pervasive issue across social media platforms, and TikTok is no exception. However, the scale and nature of the problem are challenging to quantify accurately. A recent study published in the journal New Media & Society presents a different perspective. The research revealed that TikTok’s algorithm appears to be reluctant to serve news content proactively. According to the study, “almost no evidence of proactive news exposure on TikTok’s behalf” was found.

The study employed various methods, including analyzing accounts recommended to new users and using a team of 60 TikTok-scrolling robots trained to assess video content against New York Times headlines. Out of 6,568 videos, only six qualified as news, and out of 10,000 recommended accounts, just 18% were related to news content. In contrast, the same robots, when showing interest in football, received football-related recommendations 88% of the time.

Several factors may explain this discrepancy. It could be that news publishers haven’t fully embraced the platform, resulting in insufficient news content. Alternatively, the algorithm might not recognize “news” as a coherent user interest category. According to Nicholas Diakopoulos, one of the study’s authors, definitions of news may vary, making the assessment complex.

TikTok contends that its “For You Page,” the app’s primary feed, treats news content like any other, pointing to verified news organizations with significant followings. Nonetheless, the study’s findings are striking, considering that one-third of adult TikTok users in a recent Pew Research Center survey claimed they regularly obtain news from the platform. TikTok also grapples with misinformation on various subjects, including news.

The challenge lies in comprehending an algorithmically tailored platform with millions or even billions of users. In the past, social media companies granted researchers access to study their platforms, allowing for insights and assessments. However, TikTok’s data access is limited, and companies like X are more protective of their data.

The social media giants have pushed back against the EU’s demands, asserting their commitment to tackling fake and manipulated content. However, regulators often depend on self-reported efforts by these companies, making it difficult to verify the extent of misinformation issues. Misinformation on TikTok is a genuine concern, but gaining a comprehensive understanding of the problem remains a gradual and intricate process, often occurring after harmful content has already spread widely among users.