It was my teenage daughter and her cousin who first brought it to my attention. She was texting to him on the couch, and then looked over his shoulder to read what she had just sent to him. She exclaimed, “that is not what I said!” Then she looked at her phone and in fact it was what she had written. I asked her how the message was different. She said it was mostly what she had written, but had one or two words that were different. I dismissed it as an autocorrect error-—it happens to us all. She dismissed my diagnosis and remained confident that there was something malicious going on, and she would get to the bottom of it. I didn’t give it much more thought until a couple of weeks later when I had something similar occur to me at work.
I had messaged a co-worker asking them to shift their position on a project proposal, asking them to be more in alignment with our union position, but after the proposal was submitted, they clearly had gone the other direction. I asked them why they didn’t listen to what I said, and they said they had. I checked my messages, and sure enough I had. Why would I do that? I distinctly remember taking a stance that was in alignment with our union objectives. After this occurring a handful of other times I began writing down what I said to anyone over text message or online chat. I began to keep a detailed diary of what I said to anyone online, and sure enough I began noticing that my messages were changing over time. I would type in one thing, and the message would arrive to sender transformed, and then it would also shift on my own device. My messaging software was not just auto-correcting my words, it was autocorrecting the meaning of what I was saying via my mobile phone and laptop computer.
After six months I have identified 368 separate situations where the meaning and context of what I was saying slightly shifted, which most of the time did not have much of an impact, but as with the original project proposal where I first noticed all of this, some had profound effects on my personal and professional life. 76% of these shifts occurred on mew workplace devices, with the rest occurring on my personal devices. In a work environment the shifts always seemed to be in alignment with the direction of the company I work for. In the home environment I couldn’t see enough patterns to understand where the shifts in meaning might be driven. Clearly leadership at my company had some had in shaping the algorithms that transformed my messages in a workplace context, but who was doing it in my personal life? Government? Corporations? I just can’t tell who the puppet master might be in my personal world. I would have to keep logging these subtle shifts in how I communicate, and gather more data that would help me better understand what was going on.
At first I was blaming Facebook and Google for doing this, but it was also happening on SMS, and other channels that aren’t controlled by these tech giants. Moving forward I trust no device or platform. I log all of my digital conversations. I’ve even become suspect of other digital elements like photos, documents, and videos. Could these ubiquitous digital elements in my life be manipulated? The entire experience had a gaslighting effect on my reality. I almost don’t even trust myself completely. If it wasn’t for the hand written journal that I do not let out of my sight, I wouldn’t entirely be quite sure of what is going on. I haven’t even talked with anyone about this because I don’t want to be considered crazy, and having lost my mind. Thinking back, I remember autocorrect becoming pretty annoying across my digital devices, but then it just became normal for there to be mistakes, and over time machine learning seemed to be smoothing things out and was getting better at suggesting what was “needed”. Then the autocomplete got more precise at finishing my entire sentences. Could algorithms now be autocompleting the entire meaning behind what I am trying to convey to others in this digital world we’ve created for ourselves?
