AI Image Generation and Manipulation

Generative AI has created palpable excitement nearly everywhere. But is it perfect? Um, of course not. For example, it’s not clear that AI can convincingly generate professional-grade copy yet, although it’s rapidly making advancements. Perhaps further along is generative AI’s value in the design arena, where the design community is learning to experiment with and embrace the tools in order to be on the leading edge.

Indeed, many folks with decades of machine learning and robotics experience are excited and even enamored with generative AI’s ability to support the overall design process, especially its particular knack for “seeing” a partial image—of an object, for example—and completing it automatically.

At CMD, generative AI has allowed our experienced creative team to spend less time pushing pixels in production and visualizing concepts and more time crafting original campaigns that stand out.

The powerful combination of great designers and smart AI tools

Our creative team has been leveraging many AI design tools in our day-to-day. The bedrock software has long been the Adobe suite of tools, so when Photoshop and Illustrator released generative AI features within those programs, our workflow naturally benefited.

Where it previously took hours or even days to manipulate or retouch an image, or where a reshoot may have been required, AI has made it possible to expand images on any side to create extra breathing room, clean fingerprints off a surface, and up-sample or replace an object in minutes.

The beauty of these programs is that they are built for intensive editing, so when AI provides a fix that is often not quite right, our designers can fix it, which is one of the big reasons Adobe sits apart from so many AI design apps on the market right now. Another key factor: Adobe earnestly attempts to compensate artists for their work and produce rights-cleared imagery.

Concepting with more horsepower

On the hunt for the next big idea that will break through and make our clients successful, our teams are coming up with original concepts that in the past were quite difficult and time-consuming to represent visually.

While AI struggles to generate truly original concepts, it does very well creating images that help support our ideas and make concepts come alive. This support enables us to communicate visions faster, which helps our clients get campaigns to market faster.

Avoid these pitfalls if you use AI for commercial design

If you decide to use AI for commercial design, some words of caution:

• Check the copyright. Images created by many of the most popular AI programs infringe on copyright and intellectual property. Ensure your prompts do not ask for copyrighted work that you intend to use commercially. If you are generating imagery that includes elements that should be licensed before use, do your due diligence to get clearance before use.
• Check the quality. AI can create what appears to be finished, polished imagery—but if you look closely, there’s almost always something not quite right about it: extra appendages on people and animals, objects melting together, perspectives that are physically impossible, and more fun surprises that even the keen eye can miss. An experienced designer can and should own the process and manually correct and finish images professionally.

AI in design has come a long way very quickly but also has a long way to go before much of the imagery produced is commercially viable. We see the potential for the next stage of AI to one day do even more to help support design systems, video generation, and 3D asset creation, and we’re excited to keep learning, testing, and implementing creative solutions with this rapidly evolving technology.

Read our Responsible Use Statement