One of the questions I frequently receive is, “What is the best AI tool for generating images?” 

It got me thinking about a physiotherapist who once wisely told me that “the best form of exercise is the one you actually do.” I believe the same applies to AI tools. 

For a while, Midjourney has been the leading service for generative AI solutions for images. However, as the quality of images improves with various tools, other competitive parameters such as user-friendliness, accessibility, and the ability to instruct AI without needing a long list of advanced commands become more crucial. 

In this light, it’s interesting that DALL-E is now available in ChatGPT, which many of us already use frequently. Here, DALL-E can be used without the need for special knowledge or experience as in Midjourney. 

Additionally, Adobe’s generative AI, “Firefly,” has been launched as beta version 2 and will soon be fully integrated into the entire Adobe suite. Finally, as Office 365 Co-pilot is now being rolled out, we are taking another important step towards better integration of AI into the tools we already use in our daily lives. 

Looking closer at Firefly, user-friendliness stands out the most. Firefly has a more intuitive interface and comes with prompt assistance, making it easy even for novices to use. Adobe’s new AI model is trained on twice as many images and has clearly seen a significant quality improvement from version 1. 

I am obviously not the only one seeing that Adobe is moving in the right direction. So far, Firefly has already generated 3 billion images, and its usage among Photoshop users has been 10 times higher than usual when Adobe launches new features. 

I still believe that Midjourney creates the best images, but if they fail to break out of the Discord environment or otherwise create a better user experience, I’m not sure they will have the most users in the future. 

Finally, Adobe is also a step ahead when it comes to copyright compliance. They train their generative AI model on data from Adobe Stock and actively work to make the content more transparent and credible. In conclusion, I believe Adobe is well on its way to becoming a serious player in generative AI. 

Returning to the advice of the physiotherapist, my recommendation would be to explore the possibilities in AI image generation regardless of the tool and not let potential entry barriers, such as those that may arise with Midjourney, stand in the way. 

/Henrik Ebbesen 

Group Digital Director