Review: Adobe Firefly's text-to-image

Firefly offers a suite of AI-enhancing features, but let's start with text-to-image prompts.

There are two (potentially three) significant distinctions between Firefly's text-to-image capabilities and those of MJ and other competitors. I was lucky enough to gain early access to the beta, so let's begin by discussing the most prominent feature: text-to-image.

A different standard for content

At present, Firefly is firmly rooted in transparent creative community ethics. Their generative AI is exclusively "trained on a dataset of Adobe Stock, combined with openly licensed work and public domain content where copyright has expired." This makes their tool the only one (to my knowledge) that assists creatives in generating commercially safe outcomes.