A Performance Review

AI is like your super-keen but slightly tone-deaf colleague, who somehow manages to do everything faster than you, but is still too unpredictable to introduce to clients.

And with that very concept in mind, let’s see how AI is working out for us here at Gosling: Does it have what it takes to be a fully-fledged healthcare designer?!

Strengths

With a very strict brief and some stern prompts, AI can quickly decimate a task list. Whether it’s refining, retouching or reproducing, it’s a valuable tool that helps free up designers’ time to focus on the super-skilled stuff. The thing is, while AI can take care of the repetitive tasks, and is a real pro at patterns, it can’t always nail the bigger picture.

Say your stock photography search draws a blank: AI can help create a visual from scratch, delivering something striking in minutes. But it doesn’t always hit the mark in terms of the deeper concept or emotional resonance behind the brief. It can theoretically check all the boxes, but it can’t really ‘think’ outside of them, and so needs to be closely supervised by an experienced designer.

AI’s also a bit of a perfectionist. It loves a crisp, clean line and high-gloss finish. That’s great, until you need something with a bit of grit, something that doesn’t look like it’s been churned out by a machine-made algorithm. Want a bit of soul? You’re going to have to call that designer back over.

AI can quickly decimate a task list. Whether it’s refining, retouching or reproducing, it’s a valuable tool that helps free up designers’ time to focus on the super-skilled stuff

Weaknesses

While AI can be a great thought-starter for ideation, it lacks the subtle intuition that humans bring to the table. Design is about more than just filling a blank space or matching some pretty colours – it’s about creating something that feels right. Particularly in healthcare, where we can be tackling difficult topics, or trying to make high-science seem approachable, and where social intelligence and an innate understanding of cultural context really matter.

Our eager AI colleague also struggles with maintaining a consistent visual voice across different materials. So, while it can produce more designs than the rest of the team put together, it often plays fast and loose with brand guidelines and needs reigning back in to ensure clarity.

While it can produce more designs than the rest of the team put together, it often plays fast and loose with brand guidelines and needs reigning back in to ensure clarity.

Room for improvement

Motion design is another area where AI is starting to show promise – but again, it's a bit hit and miss. AI-created images don’t have the layers and depth needed for some animations, so what feels like a short-cut can end up being limiting in the long run. Sometimes it’s helpful though. Let’s say you're editing video and need to extend a clip by a couple of seconds. Normally, you’d have to freeze-frame it, which will do the job, but won’t win you any awards for sweeping cinematic artistry.

Enter AI. Some tools are now able to generate additional footage by essentially stretching a clip, then filling in the gaps with movements and expressions. Sounds magical, right? Well, it’s getting there, but there’s still a fair bit of wizardry required. Sometimes the AI-generated footage is downright eerie. It can be tiny things, like someone blinking too slowly, but they’re obvious enough to break the spell. Alternatively, it may look blurry, flickery, or just plain wrong, which is where traditional editing comes in: We can cover any blips with filter, text or graphic overlay, or even rotoscope pre-existing footage if needed.

At other times, things swing way too far in the other direction. If you’ve seen any of AI’s best ‘deepfake’ work, you’ll understand why it raises concerns. Digitally cloning someone with alarming accuracy and without their permission is invasive and misleading at best; at worst it opens the door to all manner of misuse.

Fake it or make it?

Deepfakes aren’t the only AI-capability that raises ethical concerns. There are also question marks around originality and ownership. AI pulls from vast data sets, some of which include the work of artists and designers who haven’t agreed to their work being used. Adobe, for instance, claim Firefly doesn’t infringe on third-party intellectual property (IP) rights, while simultaneously saying they’ll pay bonuses to contributors whose content was used in their dataset training. Confused yet? Then try applying for copyright for AI-generated art. You’ll somehow have to prove your whizzy new work had a human author, without any clear legislation about what ‘author’ means in the world of AI.

Performance improvement plan

So, what’s the verdict? Could AI displace any of our designers? No chance. It’s nowhere near ready to deliver on a brief without some serious handholding. However, it has demonstrated proven benefit, sowing the seeds of some creative new concepts, and helping us enhance efficiency and streamline process.

It’s nowhere near ready to deliver on a brief without some serious handholding

As AI advances, we’ll continue to integrate the best of its capabilities into our ways of working, so that our teams and clients stay at the forefront of the industry. If you’d like to discuss how we could use AI to maximise your impact and budget, contact us.

Have questions about one of our articles?

Get in Touch