Sketch and AI
We’ve been thinking a lot about what AI means for Sketch for some time now. Here’s how we might use it, how we will never use it, and what guides our thinking.
AI has been a talking point in our industry (and in our team) for some time, especially in the last few months. We’ve been thinking a lot about what it means for us, and today we’re sharing our take with you. Here’s how we might use it, how we will never use it, and what guides our thinking.
Our mission at Sketch has always been to create great tools for designers. So when we consider all the possible use cases for AI in Sketch, that’s the first thing we think of — how can this be great for designers?
We’re not ready to make a move with AI just yet — for reasons that will become clear. However, we wanted to share the principles that will guide our approach when that time comes. Let’s talk about them!
AI might be in our future, but it won’t be generating designs
Everything we do starts with a respect for design as a practice. After all, as designers ourselves, we started by making tools for our own needs. It should be no surprise that we don’t want to adopt or build generative AI that creates designs.
With that said, we do see an opportunity for AI to aid designers in their daily workflows. We’ve all lost time to tasks like renaming layers, organizing files, isolating images, or finding that one specific Symbol deep in a large Library. Any steps we take would be in this direction.
You should own your work — that will never change with us
We feel strongly that AI should not exploit the work of you or anyone else. We will never consider it in Sketch. Not now, not tomorrow, not ever. Training an AI on your designs, for the benefit of everyone else, is not for us. Not even in an opt-out or opt-in scenario.
If we ever do see a use case that would require training an AI model on designs, our stance is clear: we would only ever do it on your device, privately, for your benefit only. More than that, you would be in full control of what happens. Right now, this is not practical nor possible, and we’re fine to wait until it is.
How about using third-party models? We can be clear here as well. We will not use them unless they can prove they have permission from their data sources and disclose them transparently. To our knowledge, such an ethical model does not exist yet — and, once again, we’re comfortable with waiting until it does.
Our approach to your data has been this way from the very beginning. You can work locally and privately. You can take your files anywhere. We don’t track what you do in your documents. No-one can view your designs (not even us) unless you let them. Why should AI be any different?
Right now, we feel that AI has not matured enough from an ethical perspective for us to adopt. That said, we remain excited to see how it will evolve and how it can one day be a part of Sketch.
We hope it’s clear that these decisions are guided by values that matter to us — prioritizing your work as a designer, your privacy, and your ownership of data. If we can’t do things in a way that aligns with our principles and values, we’d rather not do them at all.
One thing we do know, is that as AI evolves, we’ll be talking to you about it. We’ve already had some great discussions on the subject on our forum, and we’d encourage you to continue sharing your ideas there. We’re listening!
Yours,
Emanuel Sá & Pieter Omvlee
Co-founders, Sketch
