Misinformation on Digital Platforms, Politics, and Algorithmic Bias — The Risks of Exclusion

Value of the Information & Strength of the Argument

The lecture provides a valuable overview of the intersection of misinformation, algorithmic bias, and African languages, highlighting a critical but often overlooked issue. It effectively uses a compelling anecdote to engage the audience and supports its arguments with references to academic studies and reports. However, the argumentation is somewhat general and lacks deep critical analysis. The speaker presents a clear narrative but does not delve into counterarguments or complexities. The value lies in raising awareness and proposing practical mitigation strategies, though these are presented at a high level without detailed implementation plans.

Scientific Rigor, Source Quality, Title Accuracy

The lecture demonstrates a reasonable level of scientific rigor by referencing several academic papers and reports, such as those by Quelle et al. (2023), Cariolle (2024), and the CIPESA report (2025). However, the sources are not cited explicitly during the talk, and the speaker does not provide direct quotes or data. The title accurately reflects the content, focusing on misinformation, politics, and algorithmic bias in the context of African languages. The lecture is well-structured and aligns with the title, though it could benefit from more in-depth exploration of the cited sources.

The title accurately reflects the content, focusing on misinformation, politics, and algorithmic bias in the context of African languages.

7/10

The lecture is based on academic references and reports, but lacks direct citations during the talk and relies on anecdotal examples. The speaker is an academic, but the content is presented as an introductory lecture with limited depth.

Markers derived by PSI from the transcript: the creator did not define chapters.

• 00:00Introduction and course context
• 01:00Anecdote about misinformation in rural Kenya
• 03:00Definition of misinformation and disinformation
• 05:00Research on linguistic homophily and misinformation spread
• 07:00Political dimension and examples from African elections
• 09:00Algorithmic bias and language exclusion
• 12:00Risks of exclusion: information deserts and health crises
• 15:00Mitigation strategies: datasets, models, moderators, audits, policy, education
• 18:00Conclusion and discussion questions

• CIPESA - State of Internet Freedom in Africa Report 2025 — Referenced for AI-generated deepfakes during elections and policy recommendations.
• Paradigm HQ - Algorithmic apartheid? African lives matter in responsible AI discourse — Introduced the concept of algorithmic apartheid.
• African Leadership Magazine - Why Africa must confront the fake news economy now — Cited for the Uganda Ebola outbreak misinformation example.
• TRT World - Why African languages are getting lost in the AI revolution — Referenced for tokenization costs and language exclusion in AI.

• Quelle, D., Cheng, C., Bovet, A., & Hale, S. A. (2023). Lost in translation: Multilingual misinformation and its evolution [Preprint]. arXiv. — Supports the claim that a third of misinformation crosses linguistic boundaries.
• Cariolle, J. (2024). Misinformation technology: Internet use and political preferences in Africa. World Development, 177, 106358. — Supports the claim that misinformation reduces trust in government.
• Shahid, F., Elswah, M., & Vashistha, A. (2025). Think outside the data: Colonial biases and systemic issues in automated moderation pipelines for low-resource languages [Preprint]. arXiv. — Supports the claim that moderation pipelines misclassify African language content.