🌱 Algorithmic transparency
Objective
Understand what data platforms extract, how it is processed, and what influence algorithmic systems have over information flow — at the individual level (distraction, personalisation) and the societal level (narrative shaping, trend formation, information trust). Examine regulatory responses and what genuine transparency would require.
Why this topic
Personal experience of losing intentionality on platforms — arriving with a purpose and being pulled elsewhere. Broader concern about how algorithmic systems shape what information spreads, to whom, and what that means for public knowledge and trust.
Working thesis
Algorithmic opacity is not a transparency gap, it is a power structure. The volume and type of data processed — engagement signals, content consumption, behavioural patterns, inferred intent — is so vast and complex that no individual can meaningfully comprehend what is being collected or how it shapes what they see. Companies that control this black box do not just influence user experience; they control information at a societal level.
Key tensions
Full algorithmic transparency may be technically and commercially impossible — models are trained on billions of signals and outputs are not human-interpretable even to the engineers who build them. Explainability and transparency are different problems. Also: transparency alone does not restore agency if users cannot act meaningfully on what is disclosed.
Open questions
- What exactly do the EU Digital Services Act and similar regulations require platforms to disclose, and are they being enforced?
- How does this vary across regions — which countries lead on user data protection?
- Does Germany's GDPR implementation set a useful benchmark?
- Do these rules extend to streaming platforms?
- What does this mean for how future platforms are designed?
- What categories of data are actually being collected and processed — behavioural, inferred, contextual?