As part of our team inquiry into YouTube as an educational technology, I’ve been exploring the English with Emma channel and how algorithm-driven content delivery impacts learner autonomy and inclusivity.
My focus is on a dual critical issue: (1) the role of YouTube’s algorithm as an educational gatekeeper, and (2) how sociolinguistic characteristics (such as accent, race, gender, and perceived nativeness) influence which English instructors gain visibility and credibility on the platform.
Through my experience so far, I’ve noticed how YouTube’s algorithm tailors content recommendations after only a few video views. While this personalization offers convenience, it also narrows the learner’s exposure to diverse linguistic models. For ESL learners seeking to develop well-rounded communicative competence, this presents a serious pedagogical limitation. Rather than scaffolding a progressive learning journey, the algorithm seems to optimize for retention and watch time (not language development) (Tufekci, 2015).
A striking pattern is the visibility of “native” English speakers who fit into a very particular profile (usually white, Western, and female) many of whom teach with Standard American or British accents. This reinforces what Jennifer Brook (2011) calls “native-speakerism,” a bias that privileges certain forms of English and sidelines others. For learners across different cultural and geographic contexts, this presents questions of linguistic representation and psychological ownership of English.
I believe this issue is important because it connects critical discussions around digital equity, learner agency, and the sociopolitical construction of language authority. If educational technologies like YouTube are shaping the choices, voices, and norms learners engage with (often invisibly) then we need to better understand what pedagogical power is embedded in algorithmic systems. As Selwyn (2010) argues, critical inquiry must go beyond surface-level affordances and interrogate the deeper social and political forces at play.
I invite others to consider:
- In what ways do learners lose or gain autonomy when their ESL content is curated by a platform like YouTube?
- How might content creators subvert or challenge dominant norms to promote linguistic diversity?
- Are there examples of YouTube channels that actively resist algorithmic homogenization and amplify underrepresented accents or identities?
Your thoughts and suggestions, (especially on tools, theories, or frameworks to help explore this further) would be incredibly valuable as I continue to develop my individual learning plan and final paper.
References
Brook, J. (2011). The affordances of YouTube for language learning and teaching. HPU TESOL Working Paper Series, 9(1-2), 37–56.
Selwyn, N. (2010). Looking beyond learning: Notes towards the critical study of educational technology. Journal of Computer Assisted Learning, 26(1), 65–73.
Tufekci, Z. (2015). Algorithmic gatekeeping. Scientific American, 313(5), 78–81.
Thanks for sharing this very thoughtful reflection Alex and for teasing the idea of what happens when the YouTube recommender engine starts to become the “teacher”, curating the content and the learning pathway of the learner. One doesn’t need to dive headlong into complex AI systems to start to see some of the issues that might be encountered when people talk about the potential of “personalised learning platforms”; we can see (and have been able to to for some time) some of the challenges in platforms like YouTube already. I know you are looking at the implications for ESL students in particular of being exposed (or not) to a diversity of english speakers, and to stereotyped ideas of who the “native” english speakers of the world are, and whilst her work isn’t as specific as this precise use case, if you haven’t come across Ruha Benjamin’s work on the intersections between technology and race, you might find her thought-provoking. If you want a little more Neil Selwyn, then his book “Should Robots Replace Teachers?” is always worth a read, and I’d suggest thinking about Tim Fawn’s reading for the class and the idea of entanglement as I think you’ve expressed a good example of that kind of entanglement in this post.
Hi Alex,
I really liked how you framed the algorithm as an educational gatekeeper, that’s a really interesting way to put it! It really made me reflect on how easily the algorithm shapes what learners see and don’t see, especially when it comes to accent, race, and perceived “nativeness.” Your point about this narrowing learners’ exposure to diverse linguistic models really hit home, especially when thinking about what “good” English is perceived to be.
I also liked that you tied this to broader issues of learner autonomy and representation. It is thought provoking to think about how much control learners actually have on these platforms. I’d be really interested to know if you’ve come across any creators who are actively pushing back against this homogenization, or who work to highlight different varieties of English.
Recommendation algorithms are quite a double-edged sword. As long as content is served through a particular platform, it is going to move with the tides of the algorithm (which can change frequently). For a content creator with monetised videos, having more of their videos suggested to a user is the goal. Video-titling strategies, content length, thumbnail images—there are a lot of creators vying for the top spot. A content creator is not competing with just other creators but with the algorithm’s decision-making. You’ve asked good questions. If there are YouTube channels that actively resist algorithmic homogenisation, how could we even find them?
This is a fantastic angle from which to critically examine video learning platforms. Greater transparency into YouTube’s algorithm could help reveal both the positive and negative consequences of how educational content is recommended, particularly in terms of promoting unbiased and meaningful learning experiences.
There are significant implications if the algorithm prioritizes ad revenue over educational value. From a business standpoint, this may be justifiable; however, it raises an important question: Does the responsibility lie with the platform or the consumer to ensure a shared understanding of how content is curated and distributed?
You’ve posed some excellent questions that invite deeper investigation. While they may not always yield statistically significant results, reframing them to emphasize broader impacts could help identify meaningful behavioral patterns on the platform.
Here are your original questions, reframed to focus on impacts rather than specific examples:
– What are the implications for learner autonomy when ESL content is curated by algorithm-driven platforms like YouTube?
– How can the actions of content creators influence dominant linguistic norms and promote greater diversity in language representation?
– What are the potential effects of resisting algorithmic homogenization on the visibility of underrepresented accents and identities in online language learning spaces?
Like you, I’ve approached this inquiry with a degree of skepticism, especially when viewed through an educational lens, given the potential for serious repercussions. That said, there are also compelling positives to uncover, which can deepen our understanding of how these platforms influence learning.
For example, YouTube has contributed to the democratization of teaching, providing opportunities for individuals who may lack access to traditional academic institutions to reach broad audiences and even generate income. What other potential benefits might you explore to balance your inquiry and present a more balanced perspective?
Great perspective, Alex; your inquiry about the impact of the upregulation of specific English speaker profiles would be an intriguing PhD project. It would also be interesting to understand how the decisions and actions of the top-rated speakers have enabled them to achieve “market” dominance. Some questions that came to mind:
• Are the speakers applying intentional behaviours or speech patterns that are likely to achieve greater exposure?
• Are they marketing their YouTube channels through other means? (e.g. podcasts, advertising, etc.)
• Are they including controversial or tangential topics to leverage the algorithmic behaviour that tends to promote controversial topics to sustain attention?
Interestingly, when your team delivered your group presentation, I looked at the English with Emma channel and the top 3 videos on my “For you” feed were not focused on English-speaking skills:
1. “Should I move to Canada?”
2. How to use ChatGPT to learn a language
3. 5 English Idioms to MOTIVATE and INSPIRE
Of her highest viewed videos, most focused on IELTS skills, which, to your point, diminishes the exposure to other dialects and homogenises the English language to conform to a standardized language model. Within my project, I encountered similar problems with the commercial elements of Coursera, whereby Coursera leveraged its massive student datasets to promote content that is more likely to secure a paid subscription or where learners might be more successful in completing the recommended certification, degree, or program. The decline of individual agency and the use of behaviour engineering is concerning, particularly when the intentions and motivations underpinning the algorithms are not transparent to the end user.
And I agree with your remark regarding Selwyn and the need for deep critical analysis. YouTube’s content creators are political, social and economic actors, and the influence and exposure they have are producing unintended consequences and externalities.