I chose to use the prompt “As a society, we are…” and I was quite entertained by the sentence that predictive text strung together:
“As a society, we are going camping with our parents to get them in our small group discussion to make them feel free and we will need them for our future and we will definitely be going camping with them again soon.”
Based on my generated statement, you may be able to guess what my favourite outdoor activity has been this summer! I find it both funny and interesting that my predictive text statement reflects topics and themes relevant to my life right now, such as camping and group discussions. I suspect that many of my predictive text suggestions are influenced by my most texted about topics, and words that I frequently type into my phone. For example, a lot of my camping planning is done via text, so “camping” has been a very commonly used word for me lately. I am in another class right now where there are both small group discussion forums and large group discussion forums, hence why I think “small group discussion” made its way into my generated statement.
Despite how often I text or type about these topics, I was still surprised to see them pop in my predictive text in this context. When I chose my prompt, I was expecting the sentence to go in a much different direction than it did. The sentence took on a life of its own, and was unrelated to the first part of the sentence: “As a society, we are.” Although my sentence was admittedly a little out there, parts of it were quite profound. I was especially struck by the comment about making our parents “feel free” and that “we will need them for our future.” In a way, this is related to the original prompt, and it offers a sentiment that would not have originally come to my mind when considering the prompt.
The generated statement sounds rather clunky compared to how I normally write, with lots of unnecessary “and”s and “them”s thrown into the mix. It also jumps around from one idea to another in a fairly nonsensical way. Despite the lack of flow, this sentence still, in some ways, sounds like me. One aspect of the statement that stands out to me in particular is the word “definitely.” This is a word I am definitely guilty of over-using. 😉 The inclusion of this word in my generated sentence once again leads me to believe that the predictive text suggestions are influenced by the words I frequently type. However, the combination of words, ideas, and topics do not reflect how I would fill in this prompt if I were not using predictive text.
My generated statement is a little all over the place, so I don’t think it fits neatly into any one textual production category. Due to the colloquial nature of my generated sentence, I could see it belonging in a blog or magazine. Although, it would need some revisions first! The lack of commas and periods in predictive text would make it difficult to produce an academic sounding sentence, as it seems that run-on sentences are inevitable. I wonder if the predictive text algorithm is influenced by the phone application that people are typing their sentence into. I typed my sentence out in the Notes app on my phone, but I wonder if I would have gotten a different result if I was typing in the Gmail app or perhaps even Twitter? Something that also comes to mind is how colloquially people typically type on their phones compared to their computers. If there was predictive text in Microsoft Word, would it produce more academic sounding sentences due to the nature of the medium? I suspect that the medium does indeed play a role.
Predictive text is a convenient tool for speeding up communication, as it can reduce the amount of manual typing you have to do to send a message. This tool is likely best used for common phrases, such as “How are you doing?”, as these types of predictable sentences align best with the functionality of predictive text. I would be very surprised if anyone truly used predictive text to produce full sentences without experiencing issues or errors, as the functionality of predictive text is still so limited. In the future, however, I wouldn’t be surprised if predictive text evolves and becomes more in tune with how individuals actually talk and text. For predictive text to become more accurate, it will have to do more “listening” and snooping to create an advanced, personalized algorithm. Inevitably, this brings about issues related to privacy, bias, and censorship. Perhaps this will even become a sphere where advertising is possible. For example, typing “Let’s order…” could lead to a myriad of take-out restaurants popping up. Considering we already face targeted advertisements on social media based on our Google searches and conversations, I would not be surprised to see predictive text join the party.
Beyond privacy issues, the increasing accuracy of algorithms behind predictive text could lead to various facets of society relying more heavily on these technologies. It could potentially be dangerous to count on these algorithms always being accurate, as there is certainly significant room for error in algorithms that do not fully understand the many nuances of human communication. When considering these potential issues in the context of politics, business, or education, the potential for error becomes more dire. For example, the use of algorithms in politics may lead to society feeling they must question all communication from governing bodies, or, potentially worse, asking no questions at all and taking all messages at face value. It seems to me that issues pertaining to bias, privacy, authorship, and censorship are unavoidable, and that there will likely always be a divide between the voice of a human and the voice of AI and algorithms.
I will leave you with my honourable mention from the predictive text generated sentences I played around with:
“Education is not about the size of the chicken wraps but it’s not too bad for you to get it done.”
(Can you guess what I had for dinner recently?)
