I have a confession: I’m a Creative Director and I’m embracing AI.

Rather than something to fear, I think AI can supercharge our creative work. At The Frameworks we’re already seeing the very real benefits of practical, small AI projects with our clients.

One of those projects is AI-generated imagery. Programmes such as DALL-E, MidJourney and Adobe Firefly allow you to create imagery from scratch. I’m experimenting with AI to create original, lifelike images for clients – and the results are shockingly good.

Practical magic

Through my carefully-worded prompts, I can conjure up the perfect image for clients. Gone are the days of trawling stock image libraries; I can craft the exact scenario I need in minutes.

And the image quality at this speed – and price – is staggering. AI-generated imagery is now at a level where it can be used effectively in campaigns for any industry.

Take our work with UST. When the team approached us to design a microsite for UST’s AI services, solutions and platforms, it made sense to use AI to help shape the visuals. I quickly generated imagery for the site using Midjourney, directing realistic yet engaging scenarios for notoriously difficult-to-demonstrate areas like healthcare and manufacturing.

Don’t get me wrong: there will always be a place for photographers capturing real shots. But 100% customisable AI imagery is helpful when the client wants to go live “now” and budgets simply aren’t there. For smaller businesses trying to compete with bigger rivals, it offers an exciting opportunity to level the playing field.

Let’s get (really) personal

Personalisation is the goal for most marketers as it typically leads to better engagement – and soon we’ll be able to make it happen, at scale. I think AI will make tailored visual experiences for each market a standard practice. 

The speed at which AI tools can generate imagery gives us more time and flexibility to create, based on individual consumer preferences – or even serve dynamic content based on user data. 

AI will make optimising ad campaigns a lot easier, too. There’s often talk about A/B testing, but finding the right images to pull it off can quickly snowball into a huge task. With AI, marketers will be able to create the perfect scenarios for testing and determine imagery effectiveness on every campaign.

Powered by imagination

I should point out that image generation still relies on my human creativity – and a lot of prompting – to get the right images in a brand’s style. It’s fast, but we’re not in "at the click of a button" territory yet. I think of it as "co-piloting".

Crucially, I still feel like I’m flexing my creative muscles. Working out how to generate the image I have in my mind demands thought, imagination and a lot of practice. Generating AI imagery requires more than just writing prompts; depending on the programme you use, it can also include tweaking settings, training it to mimic a specific style, feeding it image references – and testing and refining all of the above.

I’ve had to learn a whole new skillset, which has been challenging, but it’s so rewarding to turn what I’ve been visualising into a picture that meets a campaign’s goals.

Ethics first

We’re set to see even deeper AI integration that will predict customers’ preferences and deliver hyper-personalised visual content. This could even stretch to AR and VR experiences for immersive marketing.

But as we head towards more personalised, reactive marketing (you can read my thoughts on that here), we need to be careful. Transparency around the use of AI-generated content is critical; in August 2024, the EU AI Act made it a legal requirement to watermark content generated with AI tools.

There’s also the increasing risk of digital manipulation and deepfakes – just take a look at this chaotic, meme-worthy video of chef “Gordon Ramsey”. In the hands of bad actors, AI could be used to create increasingly realistic images and videos to impersonate business leaders, for example, which could damage a brand’s reputation or make fraudulent behaviour harder to detect.

Many governments are already looking to regulate against deepfaked content, but organisations should get ahead of this risk and include it as part of their cybersecurity strategy.

We also have to consider the ethics around digitally creating images of minorities and people with disabilities. AI learns from data and will mimic any bias that a data set has; this demands careful human oversight. And there’s the strong argument that we should show real images of these groups of people, who are already under-represented.

AI’s bright future

Today AI can realise any idea as an image or video. That’s enabling me – and my team – to be even more creative, faster.  

I can’t wait to see where AI takes us next, especially as the quality of its output improves every day. So, if you’re a marketer looking to embrace AI-generated imagery, let’s talk.

‍