AI is a hot topic at the moment with OpenAI training ChatGPT and causing shockwaves in the industry. My business partner Patrick Cunningham has covered ChatGPT in his latest blog.
It shouldn't be a surprise that an AI chatbot has been able to become highly intelligent and mimic human response. 20 years ago I wrote my degree dissertation on the blurred boundaries between virtual reality and reality. The human race has been on a mission to try and recreate and improve our own world and minds using machines for a long time. Forbes date mechanical learning all the way back to 1308. Digital computers have been around since the 1940s. As technology evolves the output improves and there are more ways it can speed up or replace organic processes.
With my interface designer hat on, I’m still making my own design decisions today. This includes creating design systems, layouts or conducting UX research. How long before it’s normal for these decisions to be guided by AI or at the extreme owned by AI and the code flawlessly published? I wouldn’t panic just yet, we’re not all hanging up our design tools tomorrow. Actually, I think the opportunity is presenting itself to work together with AI and produce unimaginably good websites and apps. Our best work yet will be AI assisted.
To gather my thoughts on the impact it might have I’ve listed out some pros and cons.
Pros
- Processing complex datasets and displaying the data
- Layout decisions to improve conversions
- Ideas for new app features
- Transcribing speech (AWS Amazon Transcribe)
- Designing advertising banners
- Creation of persona profiles
- Compelling content strategies with perfect grammar
- Image sourcing and editing
... it's more difficult for me to think of an aspect of interface design that couldn't be benefited. The amount of time saved would be incredible and the implementations are endless. AI would quickly become an indispensable cog in the production cycle.
Cons
- Communicating the AI output and justifying the reasoning behind any decisions to clients and the team
- Integrating AI seamlessely into existing processes
- Human error with the initial input into the AI could acheive the wrong results
- Biased output due to problems in the way the AI was initially built
- Managing underperforming AI generated work
These are all big potential challenges that will exist with humans operating AI to offer digital services to other humans.
One final thing, I’ve also asked ChatGPT what it thinks.
