In the last two blogs, I explored how AI, specifically Chat GPT4, could support content design. From understanding its capabilities to testing them in a mock experiment, I’ve seen where it shines.

And where it falls short.

This part in the series takes a step back to reflect on the journey. What did I learn? What should teams keep in mind when using AI for content? And, most importantly, why does the human touch still matter?

Where teams have no content designer at all it is probably better than nothing 

In an ideal world, every product and service would have a perfectly balanced team working together. In reality – whether it is due to a lack of resources, budget or time – it’s not always possible to have every skill set at your disposal. This experiment has given me some confidence that AI could help to produce (and iterate) a helpful level of user-centred content when a content designer is not available. 

We need to be open-minded and pragmatic; while there are still elements of Chat GPT4’s answers that I would personally change or alter – it improved the existing work a lot. Even if it only helps to move the dial a few clicks in the right direction it could affect the overall usability and help users with access needs to do what they came for a bit easier. 

Just to stress – content designers really do a lot – so where possible you should be fighting to get one into the team as early as you can. A human one that is. AI is not a substitute for your content designer but, when you do not have access to one, large language models like Chat GPT4 can help to make things more user-centred than they might otherwise be if you ask it the right questions. 

It’s also worth noting that naming a service is not as complicated as having to consider a complex user journey – and how content placed at the start of a journey can have a huge impact on the remainder. This is not something that AI can save you from if you’re lacking a content designer. 

Useful when in the hands of the right practitioner and possibly troublesome when it is not 

Throughout the experiment, I was fairly impressed with the answers Chat GPT4 gave and in particular the iterations it made when requested. However, it should be taken into consideration that I was asking the AI very specific questions as a seasoned user-centred design practitioner who has worked on a lot of large-scale services. In reality – if the team did not have a content designer then someone else could be asking the questions and they might not know what they don’t know.

In other words – had the questions been framed differently or did not link to specific guidance it is possible that the answers would have been different and not as user-centred as the versions I saw in my experiment. 

I still think it’s more helpful than unhelpful – but this is something to keep in mind. If a team were to use Chat GPT4 or another similar AI tool without a content designer present – it might be best for the UCD practitioners to collaborate with the product owner, business analyst, technology lead and a relevant senior responsible owner (SRO) to help form the questions.

This could help save content designers time in the early ideation phase 

AI tools like Chat GPT4 are powerful and are capable of doing a lot. There is real scope to use AI in the early stages of ideation or problem-solving to help create strong foundations from which designers can work. In this experiment, we looked at re-naming a service. 

It took me a few minutes to think about how I should ask the Chat GPT4 to help. I knew the resources I would need to reference and I had a good idea in my head of what some candidate service names would look like. 

Even though I find that inception fun, rewarding and good to get my head into the problem space – there could be a real-time saving in using AI to help get the basic building blocks in place. Within minutes it had produced some reasonably good draft service names and had provided a rationale for the suggestions. A content designer could then use this as a starting point to apply their experience, expertise and additional nuance by way of the context of being on a delivery day to day. 

This isn’t to say Chat GPT or AI can or should do all the work of a content designer – but what it can do is save a designer time in the initial preparation stages leaving them more cognitive capacity to tackle the harder problems. 

Needs to be used ethically 

Anyone working in the AI space knows how important questions of ethics are. LLMs like Chat GPT4 are powerful and have a lot of potential, but they need to be used with caution.

These tools produce their answers by drawing on vast amounts of data, which can include information that’s biased, inaccurate, or unclear in its origins. In some cases, there may also be questions about intellectual property and whether content is used appropriately.

That’s why it’s essential to scrutinise the sources and perspectives that underpin the AI’s responses. This is particularly critical in contexts like healthcare, where precision and ethical considerations are non-negotiable. At Hippo, we prioritise the responsible use of AI, applying our own ethical frameworks to ensure that tools like Chat GPT4 are used thoughtfully and with accountability.

The verdict?

It’s not going to replace your content designer – but it’s a powerful companion tool with real world application. 

AI is probably a long way off from replacing your content designer if it ever can. This is as much down to the nuance and the intricacies of the profession itself as much as it is about the ability to utilise soft skills like collaboration and bringing people on a journey of change. For teams with no resources at all or for time-poor practitioners who need help getting a foundational structure in place then AI can be used today to make a tangible impact and one for the better. Where you can, however, get a content designer into your delivery as early as you can.