Learning in the Loop: Behind the Feed of Algorithmic Knowledge

This module invites you to explore how AI-driven algorithms filter, prioritise, and personalise knowledge in digital education. You will learn to identify the invisible systems of curation that shape what you see online, and critically examine both the benefits and risks of algorithmic knowledge infrastructures.

What’s Changing?

In the traditional era of knowledge curation, books, academic journals, and mainstream media acted as gatekeepers, determining what was published, preserved, and widely accessible.

Today, this role has shifted to AI-powered platforms such as Google, YouTube, Instagram, TikTok, Xiaohongshu, and LinkedIn. These platforms decide what we learn and engage with—often without our conscious awareness.

Search engines rank results based on user history, engagement, and commercial incentives.
Social media algorithms amplify content that triggers strong reactions—prioritising virality over credibility.
Recommendation systems tailor content feeds, but also risk reinforcing biases, filter bubbles, and ideological silos.

As AI becomes central to how knowledge is curated, a vital question arises:
If algorithms decide what is visible, what does this mean for education, research, and truth itself?

Algorithmic systems curate what we see, learn, and consider relevant—not only on social media, but increasingly within educational environments. Yet, while we often describe these systems under a single label like “AI,” it's critical to distinguish between rule-based algorithms, machine learning, and broader AI infrastructures.

In digital education, algorithms embedded in learning management systems (LMS), job boards, or citation tools serve different goals than those on TikTok or YouTube, yet they are converging in form and function through their shared reliance on engagement data and predictive profiling (Williamson, 2017; Noble, 2018).

This convergence marks a shift in how knowledge becomes visible, as digital platforms are no longer neutral transmitters but active shapers of what users engage with. While educational technologies promise personalisation, they often replicate commercial platform logic, favouring the already visible and reducing knowledge diversity. As Gillespie (2018) observes, algorithms are not just tools, they are cultural and political agents that enforce particular ways of knowing.

This module explores how learners interact with algorithmic systems, and how algorithmic visibility, bias, and feedback loops shape their epistemic environments in both subtle and profound ways.

How AI Decides What You See

This activity will help you understand how AI-powered algorithms shape your access to knowledge, and how different platforms prioritise information based on their design and commercial logic.

Instructions

Step 1: Conduct an Algorithmic Search Experiment

Open Google in an incognito window (or log out of your account).
Search for a current event or controversial topic (e.g. U.S. tariffs, ESG investing, Brexit, AI in education).
Make a note of the first five results:
- What types of sources appear first?
- Are they academic, media, corporate, or governmental?
- Can you detect signs of personalisation or regional relevance despite being incognito?

Step 2: Compare Across Platforms

Search for the same topic on a different platform (e.g. TikTok, YouTube, Instagram, Xiaohongshu, or any learning platform you use).
Compare the results:
- What kinds of content appear first?
- What is emphasised: entertainment, opinion, or factual explanation?
- How is the topic framed differently across platforms?

Step 3: Reflect & Discuss

Respond to the following reflection prompts:

How does the ranking of results influence which knowledge becomes most visible or accessible?
What role does personalisation play in shaping what we come to believe as “true”?
In what ways does this differ from traditional knowledge gatekeeping by editors, educators, or librarians?

Submit findings on this Padlet.

(YouTube Video — Licensed under CC BY-NC-SA)

This short video introduces how AI systems inherit and reinforce bias, shedding light on how algorithmic decision-making influences areas like education, hiring, and search results. You'll learn about both the technical causes and social consequences of bias in algorithmic systems—and why this matters for what knowledge becomes amplified or suppressed.

After watching the video, reflect on the following prompts. Then, visit the shared Google Doc to contribute your response and reply to at least one other participant’s entry. The first response has been shared by the course creator to help get the conversation started.

What are the main sources of bias in AI-driven decision-making?
How do engagement-based ranking systems (e.g. TikTok’s For You page or Google search results) reinforce bias?
What strategies can be implemented to reduce bias in AI used for education or knowledge curation?

🎮 AI Bias in Action – Survival of the Best Fit

In this interactive simulation, you will step into the shoes of a start-up CEO using an AI-powered hiring algorithm to select candidates. The game, Survival of the Best Fit shows how algorithms trained on biased data can reinforce discrimination and prioritise certain profiles over others.

This exercise provides a powerful analogy for how algorithmic bias in hiring mirrors similar dynamics in digital content curation, especially on platforms like Google, YouTube, or TikTok.

Put on Your Thinking Cap

After playing the game, reflect on the following prompt:

What parallels can you draw between biased hiring algorithms and AI-driven content curation on platforms like Google or YouTube?
Consider how algorithmic decisions affect visibility, access to opportunity, and the shaping of narratives in both contexts.

Readings & Reflections

How do algorithmic systems shape what becomes visible in our learning and civic lives, and what remains hidden?

"Critical Algorithmic Literacy: Power, Epistemology, and Platforms"
Kelley Marie Cotter — Open License: Michigan State University (CC BY-NC-ND 4.0)
"Digital Inclusion Through Algorithmic Knowledge: Curated Flows of Civic and Political Information on Instagram"
Shelley Boulianne and Christian P. Hoffmann — Open License: Media and Communication via Cogitatio Press (CC BY 4.0)
"The Shift from Print to Digital Media and Its Effect on Users' Media Literacy in Critiquing the Depth of News: A Systematic Literature Review"
Meyka Septira Utami Pinem — Open License: Edunity Journal (CC BY 4.0)

In the algorithmically mediated digital world, visibility is no longer neutral or evenly distributed, it is constructed through systems designed to prioritise engagement, relevance, and prediction. As Cotter (2020) argues in her dissertation, critical algorithmic literacy is essential for understanding how platforms selectively amplify certain types of content, perspectives, and users, while obscuring others.

These systems encode epistemological decisions about what counts as legitimate knowledge, shaping not only what is seen but also what is believed to be true. For learners and educators alike, this poses a fundamental challenge: to critically interrogate the mechanics of algorithmic visibility and to develop strategies to navigate them.

This theme is echoed in Boulianne and Hoffmann's (2024) study on Instagram, where they show how young users engage in active curation to surface political and civic information, essentially learning to manage and manipulate platform visibility. However, this agency is always bounded by the algorithmic rules of the platform, which filter content based on predictive patterns. Visibility, then, becomes a negotiated outcome of user intent and platform design, raising critical questions about whose voices are heard in civic and educational discourse.

Meanwhile, Pinem (2023) traces how the shift from print to digital media has impacted users' media literacy, particularly in their ability to critique the depth and credibility of news. In digital environments where information is rapidly consumed and constantly updated, learners are often overwhelmed by surface-level visibility; what appears at the top of a feed or search result, rather than depth or reliability. Without deliberate pedagogical interventions, this can undermine critical engagement, as digital formats encourage speed and exposure over slow reading and reflection.

Together, these readings point to a central tension in digital education: the visibility of knowledge is algorithmically constructed, yet often perceived as natural or personalised. As educators and learners, we must cultivate an awareness of how these systems frame our information landscapes, shaping not just what we see but how we think. Visibility is not just a technical effect; it is a political condition.

Page updated

Google Sites

Report abuse