Justin Sullivan
Okay, let’s be honest. Using generative artificial intelligence to create large amounts of text is certainly fun, but generating entirely new images from simple text prompts is particularly compelling.
Not surprisingly, the graphics tools supported by GenAI have So it’s been a source of fascination for the past year or so. What’s interesting now is that just as early experiments with text-based GenAI tools quickly evolved from fun exercises to basic real-world applications, so too did the world of generative graphics.
At this year’s Adobe Summit, graphics and imaging leader Adobe (NASDAQ:ADBE) is evolving its original Firefly image creation toolset into a suite of business-focused applications.From content creation and localization tools to customized models and integrated artificial intelligence assistants, the new products are designed to provide enterprises with many different integration methods GenAI graphics blend into their environment.
One of the biggest announcements from the event was the launch of a new app from Adobe called GenStudio. Interestingly, GenStudio combines two sets of features that Adobe has developed over the years (its widely recognized image editing and creation tools and campaign management and reporting tools) into one application.
GenStudio can take advantage of the enhanced image creation and editing capabilities of Firefly tools, and then track the performance of the assets it creates based on metrics set by the organization.
When it comes to content creation, one of the features Adobe has added is the company’s ability to easily customize the tool by training it with about 20 Adobe images or existing graphics.
These customized mockups will allow the company to create new content and other marketing materials utilizing its signature elements and unique graphic style, potentially saving existing staff a significant amount of time. As part of the new Firefly service, Adobe has created approximately 20 APIs that organizations can leverage to help them take advantage of these new capabilities.
These custom models also help overcome a potential issue that prevents many organizations from using GenAI-powered imaging tools: copyright issues. In fact, a study by TECHnaanalysis Research shows that those who are not doing much with GenAI (70% of respondents) cite copyright-related concerns as a major deterrent to doing so.
Many other web-based GenAI tools essentially ignore brand copyrights and make it easy for other individuals (in theory as well as companies) to illegally use copyrighted logos, images, and other materials.
However, Adobe has been a strong advocate for respecting copyrighted material in its Firefly tools, and even helped launch the Content Authenticity Initiative (CAI) to ensure that copyrighted material is not used to train image-based GenAI models. (Recent testing of most other major GenAI imaging tools, such as Midjourney, Dall-E, and Stable Diffusion, has clearly demonstrated this problem.)
However, if you are the company behind a particular brand, then you will obviously want to use that brand and its associated materials when creating new content. This is exactly what the new Firefly customization model enables.
Additionally, Adobe has new customization and personalization tools that simplify things like the ability to localize content for specific markets through language translation, background replacement, and other similar techniques. The tool can also essentially automate the process of adapting content to meet the different requirements of various platforms.
Another new feature Adobe is introducing to its suite of image creation and editing tools is the integration of AI-based help in the form of the Adobe Experience Platform AI Assistant. So instead of having to figure out how to do something in a complex tool like AdAdobe Photoshop, or even speed up the creation process in a tool like AdAdobe Express – Adobe Express is designed to make it easy for non-graphics people to create impressive Impressive production design – You just tell these assistants what you want to do and they will do it. The proof will come in real-world testing of these capabilities, but from a theoretical perspective, these AI assistants are certainly a very important step forward.
In terms of content management and tracking, GenStudio has also integrated new tools to see how GenAI-generated content performs in the markets or regions targeted by new content.
This is a critical capability because while making the process of creating custom branded content easier is a good thing, the effort of creating custom content is in vain if it doesn’t work effectively in the real world. GenStudio also integrates asset management capabilities, workflow and project management capabilities, reporting tools, and more.
Finally, on the Firefly side, Adobe has also introduced an interesting new feature that allows it to quickly build images similar to the example you provided. Previously, Firefly could use a reference image to essentially learn and “inherit” the styles it could use when generating the image.
However, this new structure reference feature allows you to build a layout similar to the example image you provided. So, for example, if you have an image with an object lying on its side or a silhouette of a person, you can ensure that the resulting figure has the same basic structure.
This is a classic case of an image being worth a thousand words, as using only text-based prompts to achieve a similar structure and layout of an image is often nothing more than an exercise in frustration.
As we’ve seen from other vendors, 2024 will prove to be the year that GenAI capabilities move from fantasy to real-world production. Adobe’s new product suite and enhancements for Firefly and its new GenStudio are another great example of this phenomenon, and underscore that the world of generated graphics is also entering this more practical and efficient phase.
Disclaimer: Some of the author’s clients are suppliers to the technology industry.
Reveal: not any.
source: author
Editor’s note: Summary highlights for this article were selected by Seeking Alpha editors.
