Earlier today, I was reading an email in my Gmail inbox, and I did a double-take. The person who wrote it always writes in Spanish, but this email was in English.
And then I noticed this at the top of the email.
On the one hand, it’s empowering: she and I can keep typing away in our native languages, with essentially no barriers to written communication.
This could be useful in so many ways beyond traditional “translation.”
Think of all the places where we have unseen language barriers. For example, business people talking to product people talking to engineering people. It’s hard to overstate the communication barriers that exist in this game of telephone, and the value of being able to say “I want something that does this” and having that turn into great user stories that could then be handed to the engineers would be…huge.
That said, I have two major worries:
- The most obvious is that, while I can check the Spanish to English translation, since I speak Spanish, I cannot check the English to Swahili or English to Igbo or Businessperson to Engineer translation. In most cases the black box nature of translation won’t matter, but that’s certainly giving a lot of power to the machine with minimal / oversight. If the nuance matters, that’s worrisome. And even if it doesn’t matter, that’s giving a lot of power to whoever controls the engine.
- The number of things that will fall prey to this sort of magic — and it is magic — will grow at breakneck speed. I assume that Gmail could already have a default reply written for 80% of the emails I receive, and that their quality will keep improving. How soon until I open Gmail and when I hit reply there’s a “suggested reply” email already written out? That sounds good at first, but the “win” we’ll get in terms of convenience would come with an even bigger “loss”: ultimately it’s a person whose mind I aim to change and whose heart I hope to engage. When my email bot is talking to their email bot, two people are, quite literally, no longer communicating.
We’re already seeing the beginnings of tweens and teens trying to get away from their phones because, 10 years later, they know so much more about the downside of being tethered to their feeds.
I wonder what will put the brakes on the millions of conveniences AI provides, and what will happen to business communications over the next two years. Could it be that spending time crafting a thoughtful email to someone working for another company will soon feel like stamps and airmail paper?
Good post and I think we’ll be okay. I worked for Xerox for some 32 years and as the inventors of much of what is the internet today, we actually had patents on useful artificial intelligence or as we called it machine learning. Trial and error with human tweaking resulted in very practical and non-threatening tools such as production printing and copying devices making their own service calls late at night so things could be fixed by morning.
While the great unknown is, at times, scarey I’m hopeful and rather sure that things will be okay.