AI is an unavoidable force seemingly everywhere at the moment.
In design circles especially, discussions of ‘the robots taking over our jobs’ are more regular than in years past and the ever present technology-anxiety is at an all time high.
I decided to (temporarily) join the dark side and learn more about AI technology and how it can be implemented in the graphic design process.
The Why behind learning AI
Graphic design has always been a discipline dependent on technology. From its very beginnings on letterpress through to the use of computers to now, it’s always been about pairing creative ideas with the specific tools that can enable their communication. There’s no denying that graphic design is a technology driven field and to me, it’s my job as a design service provider to be across the available design tools and know how to use them.
I’ve been a designer for about 10 years (how time flies!) and if I want to have another 10 years in the industry, I believe it’s imperative for me to understand AI tools and evaluate how best they can be best leveraged (or avoided). There are so many designers who are sticking their heads in the sand and adamantly avoiding AI but to me it’s not going to go away and it’s only going to get better.
With this huge looming cloud of anxiety, I decided to take control and educate myself more about AI- I truly looked at it as ‘getting to know the enemy’ and getting a clearer picture of what’s moving and changing in the industry.
The Course
I enrolled in House of Gai’s Ai Branding Masterclass and was a part of the January 2026 cohort. There is endless content and education out there (lots of it free) but I was drawn to this course as it specifically targets the brand design process. It’s a mix of video and live sessions and has you complete a branding project start to finish with the use of AI. I’ve found this kind of course set-up ideally for me- it means I actually do it and have some accountability.
You can see the project (and case study) I developed for the course in full here. But as a small sample, below is a little of my project developed around a slightly medieval-themed 24-hour cafe in Brisbane’s CBD.
Going into the course, I aimed to have a totally open mind to everything. I wanted to try everything and leave my preconceptions at the door to see if I even wanted to incorporate AI into my practice moving forward. For my brain, I need to know EVERYTHING about something before I can adequately evaluate it and being an active participant in this course was important to help learn where my own personal boundaries lay.
The platforms I used in this course include
- Chat GPT
- Perplexity
- Midjourney
- Weavy (which compiles a wide range of models)
- Sora (RIP)
The Good: what it worked well for
Brand Strategy and Data Collation
A large portion of this course was dedicated to using AI in the brand strategy process and one skill it was excellent at was deep diving into competitor research and compiling that data. It’s not a one stop shop (for example, I still needed to stalk the socials of each of my competitor coffee shops to see their visuals and make my own observations) but having it scrape reviews, social media and beyond for insights was fantastic.
It was also great for creating a tone of voice and then creating copy in that tone of voice. It’s a case of ‘garbage in garbage out’ but by taking time to really knuckle down on a strategic direction, I found it could create copy at a level I couldn’t easily (or quickly). Is it a little generic and has nothing on the standard most professional copywriters could bring? Yeah, but on a lot of projects without a copywriter I can really see the benefit. When I’m in the design phase, having some ‘copy soundbites’ is really handy for creating designs that look real and help the client ‘get’ the concept I’ve created.
I’d say it was also ‘adequate’ at strategy development in areas like audience personas, brand foundations (think things like the mission, vision and positioning statement and brand values) and naming ideation. It provided some useful insights, but it required a lot of time to extract meaningful observations and they were most often pretty generic. But once you HAD those ideas, the AI models were often able to refine and expand them. I found my best outcomes were from: generating AI > reviewing, deleting/ editing and making my own additions > having AI refine its suggestions.
Just to be clear- I don’t think there’s ever a world you could completely hand over your brand strategy to AI. To me the human brain (often collaborating with another human brain) is where the magic really happens- it’s the little tangents that come from human-ing together that make brands feel interesting, unique and above all REAL. But I do think it’s a handy tool to help draft elements and add additional insights I may have missed.
A Sparring Partner
Probably my favourite part of the course was seeing how to leverage AI as a ‘member of the team’ that could question or challenge my opinions.
As part of the course, we were given a very generic brief and were charged with using AI to get clarity. Having someone (or something) ask me questions about my stance sparked a lot of ideas and took it past the point where it’s easy to stop and ‘be done’.
I found the most essential part of the process was the pushing back. AI always wants to please you and will agree with whatever you say so I found myself often pausing it and asking it to clarify or do ‘not that, but this’ style edits. As someone who works alone without a team, having a sounding board to justify my thoughts to (and question them) was really helpful- it’s less about ‘producing things’ and more about ‘helping me get to the best version of my own thoughts’.
Iterating Ideas and Positional Imagery
Obviously, generative AI is good at generating imagery. I’ve played a little in Midjourney before having a brief and some techniques on how to achieve high quality images (and then create brand consistent images in the same style) was really helpful. I would never have been able to create my gothic-stonework inspired icon set or this big library of woodcut-inspired illustrations in the scope of most projects and once I ‘found’ my style, it was easy to create a whole collection that stylistically worked together. The other strong outcome was in the brand photography and creating custom mock-ups that felt brand aligned. This made designing collateral elements (like the social media graphics) infinitely easier as I had ‘real’ looking imagery that wasn’t someone else's. This helps clients see their new branding in situ alongside the style of photography I’d recommend they use and allowed me to stress test the brand with relevant content.
In contrast- it struggled to create interesting colour palettes, typography suggestions or usable logo elements. These are parts of the design process I enjoy so while they helped me see some ‘generic responses and options’, I designed all of these elements from scratch. While the AI was great at generating assets, it’s not so great at putting those together so I still followed my usual process of mapping out collateral pieces and refining once everything is drafted. I think this is the key part to using Ai in the design process- it’s not a replacement for a designer and it will never have the taste or insight to be able to assemble things in the way the creative brain does.
The Bad: What it struggled with
Unoriginal Ideas and Being a Yes Man (... or machine)
There are no big surprises with this one. AI is an amazing tool, but it’s not great at coming up with interesting, offbeat or unique ideas. This is obviously not an issue for everyone but as my whole business is about ‘embracing and amplifying uniqueness’, I found that a lot of the ideas AI brought to me were usually expected or generic. Unless you bring interesting ideas to the table, AI is not going to be able to create anything with nuanced complexity. I found that once I had landed on my own brand concept (through my usual ideation and research processes), I was able to create interesting outcomes but it was only because I have my own unique stance and vision.
AI will tell you everything you do is amazing and on-brief so it’s important to hold on to your own aesthetic and judgement throughout the process. I think a few people in my cohort were a bit apprehensive or confused with my chosen direction (which I get, I’ve learnt I probably design in a method different to a lot of designers). On seeing the final outcome they ‘got it’, and I wonder if I had blindly taken AI’s input if that would have been the end result. To me, the brand looks like it could be all my own work which shows you really need to trust your own unique aesthetic and approach when using this technology.
Being a One Stop Shop
To no one's surprise, there’s no magic ‘design this’ button that creates great final designs. As I touched on above, what surprised (and maybe reassured me) is how much ‘traditional design’ is still used in what you’re creating. Think things like-
- Designing the logo from scratch
- Adding my own elements to mock-up images via Adobe Photoshop
- Touching-up and editing images in Adobe Photoshop
- Taking icon assets into Adobe Illustrator and vectorising them (there are AI options, but I found they lost the chunky texture I wanted)
- Designing a suite of social media graphics in Adobe InDesign to see all the elements come together
(For now), there still needs to be someone with experience, and maybe more importantly taste, driving the ship.
Limitations to what the models know
In choosing a medieval aesthetic for my brand, I thought it would be able to create imagery with NO PROBLEMS as there is so much available imagery out there. This in fact, did not happen as it seemed to struggle with “I want this but with a bit of this in this very different aesthetic” aesthetic. Initially I hoped to have a support illustration that was a ‘medieval inspired celtic knot, in a contemporary monoline way, while also being non-symmetrical’ and it just could not understand or produce this request.
Human brains can come up with interesting pathways and concepts and I found throughout the process that the most interesting visual tangents were still inspired by my own brainstorming or via Pinterest-style research. I always start the design process with a BIG deep dive into visual inspiration and that's where ideas often get sparked. I wasn’t 100% sure what my brand icon would be but in researching medieval iconography, I saw the ouroboros symbol and it INSTANTLY connected with the idea of a never closing 24-hour coffee shop. I don’t think I could have asked AI for that idea- it had to come from my own ideation and gathering methodology.
Ownership
This greatly depends on what you’re creating and where, but often you can’t ‘own’ the assets you’re creating. This is muddy ground when working with clients on finished art but like I’ve mentioned, if you’re using AI as part of the process (rather than a finished product) this is greatly a non-issue. This also isn’t an idea only in AI- designers have known for years that people can create logos in Canva but never truly have ownership (or the ability to trademark) their work, but I think it’s important to mention.
Also depending on the plan/ model you use, your work may be ‘in the public view’, which may be a confidentiality issue.
The Ugly Sides of AI
I think the brainless over-use of AI IS problematic (think asking it simple questions you could easily search for, using it to badly do your job in one-click). Generally I think it’s making us less critical thinkers and if I was in an industry that didn’t HAVE to engage with technology, I probably wouldn’t use it. I think it’s very easy to say ‘all AI use is shitty’, but to me there is a big difference in using it intentionally (and ideally minimally) in my industry where it’s becoming an expectation versus using it frivolously in my non-work life when I don’t really need to.
With this in mind- there’s also a heap of environmental and ethical factors at play. I’ve kind of landed on ‘using the tools as minimally as I can, and being mindful of HOW I use them’ rather than a blanket avoidance. Lots of these models obviously sample the work of creatives (not ideal) so I’m mindful of not directly taking inspiration from artists (ie. not feeding it imagery) and instead generating from scratch through descriptions, edited iterations and my own creative work.
I think a large part of the negative perception around designers using AI is around transparency- it’s one thing to use AI as a tool in your design process, it’s another to misrepresent work you are charging people for as ‘by you’ when it is in fact created by a machine. This is an easy one for me- I could never stand behind something I didn’t create and moving forward, I plan to be very transparent with any AI created content.
The Future
So, where to next?
I don’t think there’s any point in blindly avoiding AI if you want to be a designer in the next 2-5 years (just my opinion). I’m mindful that the design industry is changing at an obscene pace and even though I think AI as ‘a designer’ is kind of average at this point, it’s only going to get better.
OBVIOUSLY, I think it’s the ideas and strategy and brand differentiation that MAKE interesting, impactful and GOOD design. I think there will always be work for creative brains who use whatever tool is available to make their vision a reality.
To me, AI is a tool to be used- it can be used in boring ways (to automate/ start tasks I find tricky or unenjoyable) and exciting ways (help translate the wild ideas in my head into visual outcomes).
I haven’t made any concrete decisions but for now, I can see myself utilising AI for
- Competitor research (collation and analysis)
- As an educated sparring partner to help refine, challenge and improve my ideas
- To come up with ‘shitty first drafts’ of things. The blank page is a horrible beast so sometimes having something wrong or ugly or generic helps me solidify the direction I want to take (in both writing and design tasks)
- Outsourcing areas outside my discipline (eg. placeholder copywriting). I’m not going to pretend that it’s anything as finessed as hiring a copywriter BUT it’s handy for having small bites of copy to design with.
- To help communicate to clients ‘the vibe’ of something that might not exist. Sometimes I have fab ideas for a photography or packaging style that I can’t find a realistic reference for. By using AI, I can show that visual (often as part of the brand design) so the client knows what to expect and can use it as part of briefing/ hiring/ sourcing.
I know in writing this article, some potential clients (or fellow designers) will be pretty judgemental and will choose not to work or support me. That’s a totally valid stance, but just know I don’t think creativity and AI are forever destined to be enemies- I believe with a cautious, ethical and transparent approach it can be a useful piece of the brand design toolbox.
I’d LOVE to know your thoughts on AI- feel free to comment or DM me with your thoughts!
PS. Fun fact = I used zero AI in writing this blog article. I promise- it’s not all or nothing.
PPS. When I asked AI to envision 'the future of graphic design' (with not too much direction initially) it resulted in these weird, post-apocalyptic imagery of mushrooms taking over technology. So maybe we'll be... ok?
