For this task, I went with the prompt “this is not my idea of………….” on Instagram.
I have an Instagram account but I don’t use it often and I haven’t posted anything in a year or two. Therefore, I found it interesting that the text came from historical data from when I texted friends on my cell. There were phrases such as “go back to school and get stuff done first” and “what the weather looks like” are statements I have used in the past. It’s a bit scary because the words does sound a bit like me, but there is something odd about it, like the quirks from an individual personality feels flat. The predictive text doesn’t fully sound like me because it doesn’t take into account that I modify what I’m saying base on changes in the environment and who I am talking to. The algorithm just generated an output based on data (words) that I have used in the past and what works in the past should work in the future.
I think this type of predictive text can be useful if your blogging to lots of viewers on social media and since it’s to a wider audience, you can post short messages frequently that sound relatively like you. It’s definitely time efficient because you don’t have to ponder about each content. A fault with this is that you begin to rely heavily on the predictive text that you can get lazy in checking what you actually wrote to see if it is what you wanted to say. For example, there are times I have quickly texted friends and sent the message only to look at it later and notice that a word or two was not the word I wanted to use, but what the algorithm predicted.
The pros and cons of predictive algorithm reminded me loosly of the movie Money Monster with George Clooney.
In this movie, Lee Gates is a TV host who picks winning stocks. One day, an angry investor Kyle Budwell lost all of his savings on the IBIS stock when Less said it was the winning stock. The story moves on to uncover the truth and the “glitch in the trading algorithm.”
This was interesting because Lee Gates picks the winning stocks based on an algorithm that help with making decisions, predicting winning stocks, and profiling good stock buys. Some positive aspects of a predictive algorithm. However, the algorithm went wrong when Kyle Budwell followed Lee Gate’s advice and the IBIS stock went down.
Here are some of the difficulties of a predictive algorithm which were seen in the movie. Someone can build the algorithm, input the data, and set the parameter for what the output of success is. However, it doesn’t take into account that future events may change. Furthermore, it doesn’t take into consideration human error due to greed or bias. As well, the end result can be skewed based on the data provided or the fact that there just may not be enough quantitative data for an accurate prediction.
Hi Melody,
I can totally relate to what you are saying about how scary it is that our phones seem to remember phrases that we used in context and relation to others. I also found with my predicative exercise that it was linking phrases I used about my speculations on COVID to create content for what I was saying about society now.
I had however not thought about how the predicative text does not seem to account for present or future and this is indeed an interesting perspective that needs to be considered in order to improve the algorithms of the future. Also that movie looks really good and think I will track it down and watch it to fully appreciate your comments about it.