OpenAI, like many other companies and a billion YouTube videos, are pushing us to use ChatGPT or similar models for everything, including writing a thank you note to the babysitter (a lady, we suppose here). Does it make sense? Let’s try to consider it from the babysitter’s perspective (if she knew we didn’t write the note ourselves). “Thanks!”, she would say. “You didn’t even lose a second of your precious life to write the note YOURSELF”, she would think. Yes, because the time we dedicate to something is part of the value of that thing, especially if it’s time spent doing something for someone else.

Let’s broaden the perspective. Kevin Roose, in his book “Futureproof: 9 Rules for Humans in the Age of Automation,” suggests that in the era of automation and AI, instead of entering a hopeless competition with machines to do what they do better, we should focus on being more human and doing creative, inspiring, and meaningful things that, at least for now, the most advanced AI can’t do. (No, the kind message from your AI system or the happy eyes of your little robot is not genuine kindness, it’s programmed simulated kindness; it’s quicker, cheaper, and we believe it).

Time, value, and empathy are part of our uniqueness: I prefer that YOU truly remember, once every three years, to wish me a Merry Christmas with YOUR slightly grammatically incorrect note, maybe stained with sauce, rather than remember me only because an algorithm sends you a message, writes the note pretending to be Dante or Shakespeare, and prints it on a pristine, sparkling paper.

Professor White, mentioned, for example, in this post, rightly recommends using AI to enhance, not diminish, our cognitive abilities (“augmented intelligence”). Otherwise, among other things, we risk becoming more and more replaceable with increasingly sophisticated AI systems. We also risk becoming increasingly useless to ourselves.

But there’s more. In the era of climate change and the (modest, for now) efforts to reduce emissions, it’s important to consider how much CO₂ it takes to train a system like ChatGPT and how much it produces with each use (yes, because AI systems run on real, often large computers that require a lot of energy to operate). Those curious can, for example, read this article from The Guardian with the meaningful title “Why AI is a disaster for the climate”. In essence, when we’re about to use ChatGPT for something we could easily do ourselves (and make it less alienating and more meaningful), let’s also think about our contribution to climate change. It’s a small contribution, but as usual, billions of small contributions make a big impact.

In the project development framework proposed by DeepLearning.AI in the AI for Good course, it is suggested, as a first step, to evaluate whether AI truly adds value to the project under consideration. Let’s ask ourselves this question not only in a professional context, but for anything we’re about to do.