AI isn't magic. Why is it being marketed that way?
Apple Intelligence leans on mystical language that's disempowering users
When I was getting divorced seven years ago, one of the mechanisms the family courts offered to conflict-prone co-parents in an effort to make their communication with one-another more—ahem—civilized, was an app called OurFamilyWizard.
One fascinating function is now called ToneMeter, which uses embedded AI word recognition to cleanse messages of the contemptuously unproductive, or downright hostile language you will use when say, making arrangements for soccer practice:
Exclusive to OurFamilyWizard, ToneMeter is built to keep conversations between co-parents productive, clear, and civil. ToneMeter provides real-time feedback to help you shape a positive tone instead of sounding confrontational.
I remember feeling like a guinea pig at the time, as here was a then-rare technology being made available to a relatively desperate and vulnerable client-base who defeatedly accepted the intrusion into personal conversations for the greater good.
We relinquished our privacy, knowing that if our messages were corrected they were also being monitored and recorded, with the potential to be used as evidence for or against us in court. With relatively little knowledge or popularity surrounding AI at the time, I figured the word and phrase changes were being made either by an automated word replacement system, or perhaps an overseas boiler room where actual humans read my messages and suggested more pleasant or cleansed alternatives. I like to imagine the latter, and that one day we’ll meet up for coffee and my call-centre confidante would tell me all the horrors they had to clean up in his messages to me.
Yes, we were vulnerable. Yes, we were desperate. Yes, we gave up our control and submitted our language to a mysterious wizard that, corrected us. But now, AI-powered word suggestions are available to everyone everywhere. Most recently, in the text feature of the new Apple Intelligence you can choose from a drop-down menu to make your email composition more friendly, professional, or concise.
You can watch the promotional film that Apple produced in June, which is rife with images of glowing orbs, otherworldly halos, and language like magic, and Image Wand. What hasn’t changed is the inherent mystery around how these functions are contextualized. The effect of incorporating language of the supernatural to describe AI features both mystifies and relegates the understanding of how they work, making it seem like understanding these machinations is outside of our faculties. The result leaves each user feeling disempowered and blind as they’re urged to release control.
We’ve seen this in the past with hardware: Apple’s stronghold on making proprietary devices spurred a DIY movement with companies like ifixit equipping Apple customers with the education and physical tools to repair their own machines. The embedding of these functions in the new iPhone 16 will only ratchet up that reaction.
My grain of thought
Now that AI-powered features are integrated with our everyday lives, we need to look behind the curtain and know how these systems work, and who creates them. Mystifying language like magic and wizard disassociates them from the realm of human understanding, and disempowers the people who rely on them. Promoting it as something supernatural is a slippery slope, when these are actually programmable functions that can be adapted and modified to suit our uniquely human needs.