Algorithms of Predictive Text

Education is not about the fact that we are going through this process of being together. I don’t think you need to be there in person to see other benefits. There are some people who don’t have the right idea of what to do. I don’t know what the name of the game is but the only thing that matters is that we don’t go backwards.

I chose to create a micro-blog on the prompt, “Education is not about…” and used the Apple iPad textual algorithm to complete the micro-blog. Since the algorithm makes predictions on the weighted strength of the connections fostered through use,  I think that it is important to make note of what activities that I primarily use this device for. I do very little academic writing on it, instead primarily using it for personal reading, texting friends/family, and quick online queries. These activities would definitely shape the words that are suggested by the algorithm.

The statement that was generated by my predictive algorithm was similar to one that you may see on a textual product such as a blog. The usage of the personal voice distinguishes it from the other type of writing that I primarily participate in, which is academic in nature. The produced text may also show similarities with novels told from the first-person point of view, although the subject matter and writing level would be lower than those I would be reading as an adult.

This generated statement is different from the way than the opinion I would normally express on the topic. While it conveys the similar ideal of education being related to a collaborative movement forward, I would not have specifically mentioned other people since I believe that education is primarily a unique, individual experience that can be enriched through collaboration with others. This is an example of how the predictive algorithm influenced the output of the text and shaped the expressed opinion.

Furthermore, I tend to opt into the academic or formal tone when doing any writing due to my occupation as a teacher and continued involvement in formal education. If I was provided with this prompt, I would not have used any first-person personal voice. The algorithm twice uses “I” as a way to begin the sentence. This prediction would originate from my personal text messages that I occasionally use on the device, where I opt for a much more casual voice, often using personal nouns. The “voice” that the algorithm created for me sounds like a blend of the two different “tones” of voice that I use in my daily life, but is not quite able to replicate what I would respond with.

In the educational setting, I often see students using the predictive text element that is part of Microsoft Word to complete writing assignments. This often makes their writing unintelligible and the student’s comprehension of what they have produced is often compromised. With the emergence of algorithms being used in an educational setting, it is easy to see how they could support students but also take away from authentic learning experiences that students participate in as part of making mistakes and maintaining a growth mindset. I would imagine that a more complicated algorithm would also make it easier to plagiarize since it could use other academic sources as a basis of knowledge. We can also observe algorithms as an academic policing software, such as TurnItIn. This software can make it easier to maintain academic integrity, but also can reduce the role of the teacher and lower their understanding of their student’s abilities. It will be a constant struggle of education to balance progress with supportive usage of technological advances moving forward.

2 comments

  1. Braden,

    Similarly to you, I also used an Apple device to complete this task: namely, my iPhone. You raise an interesting point about not using your iPad for academic writing, but rather for more informal tasks and interactions – this certainly shapes the predictive text in a way that might differ from your other devices. Also, the use of first-person voice in a statement on education is understandably an approach you wouldn’t take if you were writing more “authentically” when compared against predictive text.

    It’s fascinating that you bring up the use of predictive text by your students to complete assignments and resulting in unintelligible work. I don’t work as a teacher, so I had no idea that the practice was so prevalent though it makes perfect sense – I’m sure students find it to be very convenient, but their ability to organically cultivate their own “voice” in their writing becomes much more difficult if they become reliant on predictive text.

    Excellent post!

    1. Hi Amy,

      Thanks for your reply! I actually went back and completed the task using the predictive algorithm on my laptop, which is what I use for academic writing. The tone and suggestions made by the algorithm were completely different from the iPad, despite being the same software, almost akin to being different people. It was a very visual representation for me of how these algorithms learn from use. I think that is also a reason that work submitted from students using predictive text can be nonsensical since it learns from their quality of work.

Leave a Reply

Your email address will not be published. Required fields are marked *