Nano Banana in Adobe Firefly, Google new ai, Gemini 2.5 Flash
Nano Banana, Gemini 2.5 Flash in Adobe Firefly. How does it effect Photoshop?
It seems like every week there is a new ai model coming out, and the headlines are often the “Photoshop killer”, “watch out, Photoshop”, and other hyperbole. But is it just hype?
The newest ai image generator to get a lot of attention is Nano Banana which can reproduce something, like a person and only change certain things like clothing or backgrounds with just a text prompt. Google announced that nano banana was a secretly theirs and is now Gemini 2.5 Flash. Adobe has already included this model into Firefly4. Ill discuss it and what I think about the future of Photoshop (spoiler, I think Photoshop is here to stay).
Let’s start with the image I’ll use. It’s a low res image of myself.
Log into Gemini.google.com to access Gemini 2.5 Flash. Click the button to upload your image.
Here I typed the text prompt; Change the picture, so that the man is flying a space shuttle.
I got this result.
A previous attempt, yielded a better result. Notice how the face is mostly unchanged while everything else changes around it. Notice in the lower right a star is added to the image. This lets viewers know the picture is ai generated. The star will show more information when you click on it.
Let’s try again, this time, we will change my Wacom tablet into a black and white TV.
Maybe we can expand the room and change the background.
As you can see it works a bit like firefly, but with better looking generated results.
Nano Banana in Adobe Firefly
Let’s go to Adobe firefly
To go into Generate mode, click on Generate and Adobe Firefly will go into the text to image workspace without having to generate anything first.
Under General Settings, click on Model to see all the available ai models. Choose Gemini 2.5 Flash image. Gemini 2.5 Flash image is the official name for Nano Banana
Let’s see what Gemini 2.5 Flash can do inside Firefly.
Here’s a selfie from Adobe Max that I have uploaded. This is at the morning keynote, before it began. There is a crowd still trying to wake up behind me, the coffee will kick in soon. Let’s see if we can make them more enthusiastic, as if the program has already started.
Type the prompt, “make the people behind the main person look happy”.
Here we see the result on the left side. Everyone is looking way more chipper. (Make sure to check over-enthusiastic crowd photos for ai use in the future).
Let’s try a different image.
Type in the prompt; change shirt color to red. It did a good job without changing much else of the image. Although, my skin does look a little softer than before.
Lets try a darker red shirt, replace the controller with a pie and change background to a bakery.
Not bad. You can see that the person (me) is starting to change a little as the amount of generations increase, but still impressive.
Watch the video at the top, to get a really good view at these examples and more.
My thoughts about Nano Banana
The results are impressive and the generated images are more interesting and usually more realistic than Firefly. One of the things I noticed when it’s compositing, it’s not really doing a great job of blending in the color and the light. In some cases, the blending is pretty good, but it seems like it’s struggling with things like depth of field. It’s struggling with light direction and color, especially the color temperature.
Maybe for a work-around, you could extract it, run it through Adobe Harmonize and make it better.
My thoughts about ai and the future of Photoshop
A big thing we’re dealing with all ai right now, is limited resolution, Gemini 2.5 Flash is only generating at about 1k resolution at the moment.
When we’re talking about a Photoshop replacement, there’s so much more that we use Photoshop for than just photo manipulation. So let’s talk about the big one. When we do an edit, even a tiny little edit, the ai is recreating the entire image. So at that point, once you’ve made even the smallest edit, your original image is gone and you’re now working on a 100% AI image. To some people, that doesn’t matter. To other people, that matters a lot.
So the way I look at this, it’s a useful tool. Photoshop is an application where you can directly touch an image and make adjustments manually using your right-side brain. Generating a text prompt can definitely engage your right-side brain, but it’s also a left-side brain activity. Deep prompting is more of a left-brained activity. Editing images with text-prompts is definitely not a Photoshop replacement. In Photoshop you can make a small adjustment without disturbing any of the other pixels. I think Photoshop is going to be around for a long time.
Gemini and all ai are tools
I do think that Gemini 2. 5 flash is really amazing, and I’ll definitely be incorporating it into some of my workflows. But it’s not going to replace all my tools. It’s another tool. And I feel that way about most ai.
Look at commercial work, it’s about getting the work done to a professional standard as quickly as possible. That may or may not involve ai, depending on your client, upon the stipulations of what you’re creating. Maybe you’re being hired to create ai-generated art, then go for it. That’s definitely a career in the future. A prompt engineer is a great future career.
But there’s definitely a place for manual work and manual skills, Even if you’re using AI, not everything can be exactly the way you want it by typing in prompts. Sometimes it’s a lot quicker to open it in Photoshop, grab that tool, make that adjustment.
Colin’s personal thoughts and predictions on ai
I believe your skills are only going to help you. A retoucher for example is still going to have a job. Instead of retouching only photographs and hand drawn art, you’re also going to be retouching ai generated pictures, because there’s still very specific things that humans want that can only be done manually. There will be many jobs where ai isn’t wanted, or is even forbidden. Editorial work will forbid ai and I doubt it will ever be allowed in most journalistic situations. Art and photography contests don’t allow it, although I’m sure there will eventually be an ai category in many. As time goes on, there will be pro ai and anti ai movements. Some countries may ban it from publications, while others will require a disclaimer. In this age of craving authenticity, I can see a flood of ai coming in and then a movement away from it. It will be widely adopted by casual users, but these aren’t normally the people who would be using Photoshop anyway.
That’s not even mentioning digital painters, graphic designers and people who use Photoshop for production work. A lot of these tasks will be ai assisted, but I don’t see it completely taking over. We will see a lot less use of Stock photography and clip art for sure. Tools like Content Aware Fill will suffer a decline in use.
Then there is the hobbyist, or the person that just enjoys working in Photoshop. For your own work, if you prefer to always do it all by hand, that’s perfectly ok, or if you choose to do hybrid work, that’s also ok. Do what makes you happy.
So never feel discouraged about the skills you’ve learned in Photoshop, and don’t give up on learning Photoshop because these skills are so useful. I believe in the future, as perhaps less people are skilled in Photoshop, these skills are going to become even more valuable. If you are looking to make a career out of Graphic Design, photography or other visual imaging, it’s also important to add ai as part of your toolbox. The flexible professional with mastery of different tool-sets is the one who will have a successful and (hopefully) long career ahead. To be the best at anything you do, mastery of your craft is a MUST. Don’t stop learning and growing.
Anyway, that’s my thoughts. I’m curious, what are your thoughts? Drop them underneath. I’d love to hear.
Colin
PS Don’t forget to follow us on Social Media for more tips..
(I've been posting some fun Instagram and Facebook Stories lately)
You can get my free Layer Blending modes ebook along with dozens of exclusive Photoshop Goodies here
One thought on “Nano Banana in Adobe Firefly, Google new ai, Gemini 2.5 Flash”
Leave a Reply
How to make a photo look like a tiny miniature model in Photoshop with the tilt shift filter. This Photoshop...
In this Lightroom 5 tutorial, you learn how to tether a camera to Lightroom so that the photos go directly...
How to use the Lumetri panel in Premiere Pro to color grade your video. This 11 minute mini course teaches...
Hi Colin.
having just watched your Banana, Firefly, Ai report, though very impressive, I certainly agree with you that this will not, and neither should it, replace Photoshop.
I think we have to bear in mind that (in my case and many others.) we are photographers, and we set out to capture an image has we see it and not to then change it completely by using some invisible genie to add something that was not there when the image was captured.
Photoshop For Ever!
Regards.
Phil Ashmore.