AI is like marmite, liquorice or cilantro. Most people seem to either love it or hate it.
There are many good things that can come with AI, ranging from simple productivity power-ups to the promise of curing all disease, all the while being a big threat to destroy the job market and more broadly — destroy humanity.
So I get both the love and the hate positions. Both are valid.
That’s why, just like that Daylight song, my position is — Oh, I love it and I hate it at the same time. Coincidently, this song is about hiding from sins, something that I will touch upon later in the “The hate chapter” below. Feel free to fast forward the song if you like. For the curious ones, the song goes like this:
Oh, I love it and I hate it at the same time
You and I drink the poison from the same vine
Oh, I love it and I hate it at the same time
Hidin’ all of our sins from the daylight
From the daylight, runnin’ from the daylight
From the daylight, runnin’ from the daylight
Oh, I love it and I hate it at the same time
After the short musical break… Back to the article…
As I was experiencing the last ChatGPT update, the first thing that came to my mind was:
WOW, THERE ARE NO LIMITS NOW.
I loved it, and I dropped my bowl of muesli on the floor in amazement.
Same as many others, the next thought was:
IT’S SO SO OVER FOR US.
I hated it, and I threw my new bowl of muesli in the garbage in revolt.
I dip in between those two positions daily.
What I love about it
Groundbreaking technology giving us limitless creative power
Couple of years ago, AI seemed like the new hype. It seemed like the food we are forced to eat (similar to the metaverse & NFT’s and NFT’s in the metaverse). It seemed as the next front(eer) / gold rush and a major new driver of shareholder value.
Even though this is partly still the case, we are safe to say that this thing is here to stay. It was magical even back then. What it’s able to do now is just simply astonishing. You either have to be 1. afraid or 2. spoiled or 3. having blind spots to see it and to still underestimate it. But, but — it still hallucinates. Sure! It’s hyped up. Sure, extremely! But nothing should take away from the fact that this is groundbreaking technology.
You’ve probably heard Amara’s Law before somewhere:
“We tend to overestimate the effect of a technology in the short run and underestimate the effect in the long run.”
I think we underestimate AI on both time horizons. Or we might already be in the future since we started using this sentence in the context of AI a long time ago, especially if you count in AI years. AI years go faster than human years.
Just have a look at what a designer can do now (NOW NOW):
Some might minimise the impact by saying all sorts of things: this is only for freelancers. This is mostly applicable to visual / graphic / motion / brand design, not for UX. Also: this is not vector, these are not high quality images and videos that can be used for high value campaigns and projects. This is not applicable in big companies… And all those similar comments, basically saying: I am not affected and I’ll never be, only others will. Mostly not true and from a position of ignorance or shortsightedness. There are plenty of videos like the one I shared above addressing those concerns and demoing already available tools and workflows for most of our creative output.
2. We are using the worst version of AI at the moment, it can only get better
The fact is — the way we work is changing. Fast. Going beyond the first generation outputs of GenAI, now we have: deep research, vibe coding, correct number of fingers, consistency, text on images etc. It’s gotten to a level where it can generate creative work that is actually usable (some limits apply). Not only for small projects and prototypes but actually useful for creative work used at scale (some limits apply).
Next up, in no time there will be no limits, there will be AI tools where we can have:
Fully and easily editable layers: not necessarily Figma type of layers but front-end that can be moved and easily edited (think Framer / Webflow meeting Figma meeting Lovable / Bolt / Cursor / Google Stitch). There will be a convergence of tools and ways of working because the list is endless and so so fragmented.
Improved importing of files and output of code: bridging the gap from design artefacts to a real product. I am mostly speaking about front-end and especially for small to mid simple web sites. Web apps requiring back-end and native apps are another ballgame and will still remain off limits to most of us for some time even with the help of AI. But we can for sure prototype them in a better way.
More predictable results: less of Russian roulette type of outputs of AI tools and vibe coding, especially related to the next point:
Reliance on a design system: UI consistency and creating designs and building prototypes with a defined system. A system that can be created with the tool to begin with.
Reliable automatic A/B testing: self optimising sites and apps that work within set boundaries and systems, all with set business or user goals.
Instant push to a server without setup of environments: anyone can push to production, no need for complicated set up upfront and while committing. Again, front-end only for us designers / creators, for now.
3. It can be used across the entire process
I’ve tested AI for strategic work early on in a project — specifically market analysis and generation of possible directions and was impressed with the (deep research) output. Adding decks in NotebookLM and summarising and querying them gave me some fast insights as well. These were by no means a final strategic deliverables but it was a valuable springboard to more serious thinking. It proved to be a good sidekick / analyst confirming some of my hunches while also opening some new doors.
Next, I can continue to use it for competitive analysis, as an assistant with research (not talking about synthetic research), create flows and generate copy before getting to the visual design.
To get to a good visual design, in the past (circa 2 years ago), I would be hunting for stock photos and videos for hours to take my design to the next level. And I was mostly limited to free or low cost stock websites so I had to go through a lot of these:
Now, I can just imagine, prompt, upload reference photos and… BAM. I can get a upgraded visual design basically for free in seconds, minutes max. I can do the same for mockups of products (and other branding mockups) and for icons (AirBnb style of whatever else that my eye desires). And all of that can easily be animated or made into a video. Few more upgrades for a small price and in no time.
From there, I can continue with AI for coding (basic front-end is where I will stop for now but others can go on further).
All of these steps, with hours, days and weeks shaven off the process. I would get further, faster.
Before you get too excited, there are still some limits. I am working as a Design Lead in a big corporation at the moment so I don’t use AI across the entire spectrum. What’s important is that me or others can use it across the spectrum if needed. There is limitless potential.
4. It can make designers generalists
All of the above can enable one powerful shift:
Our focus can shift from being a specialist to being a generalist. And if you were a generalist already — to being a even wider T-shaped one person army / creative studio. We can move at the speed and limits of our AI subscription. And whatever the speed is, it will still be faster than before.
Soon enough, everyone can be a generalist. Not just across the UX crafts (design — interaction, visual, UX, product; writing, research), but also: Clients, Devs, PM’s included. Everyone can be a designer, everyone can be a PM, everyone can be a developer. Everyone can be everyone, and we can all be generalists. We can all be creators without limits.
Jorge Valencia and Dára Sobaloju covered this topic extensively in case you need more arguments.
Whether a certain output of a generalist is better than the output of a specialist is a totally different question with a clear NO as an answer. I think and strongly believe so. I am not sure if all of our clients would agree or will care. If AI can make designers good enough in most areas, or if budget does not allow for specialists, I am afraid generalists will be the way forward. I am generalising here but I can see the tendency.
Soon enough, AI can serve as a limitless pill. It already does to some extent.
Most of this relates to our tooling, our output and our productivity / efficiency. It’s not a supplement of a good design mindset or thinking, this is where we come in. But we can come in stronger and faster.
And I LOVE the potential of that.
But it’s not all love… As the song goes: Oh, I love it and I hate it at the same time…
The hate chapter
Limitless bullshit
We now potentially have limitless power, but with limitless power comes limitless bullshit.
Limitless spam. Limitless number of tools. Limitless number of fools. Firehose of information about AI and generated from AI. Limitless FOMO. It’s happening. Heck — I might be part of the problem with this article!
We can do everything but should we do everything will be an even more important question going forward. Are we just transmitting, adding to the noise or creating value and valuable discourse?
There is a limit to how much we can consume. And a limit to how many tools we need. Plus a limit of web sites and apps we need and can use.
Who will read all this stuff? Who will use all of this stuff?
2. Limitless energy drain
All of it is a big energy drain. Both for us humans on this planet and for the planet itself. By an estimate, the AI sector can consume the same amount of energy annually as The Netherlands by 2027.
Whatever the exact numbers are, we will need more energy, more resources. AI will become more efficient but at the same time embedded in and used for more and more use cases. It might even accelerate the energy transition to nuclear. Oh, great idea and nothing can go wrong— another thing that can destroy humanity in the service of this other thing that can destroy humanity! Easy solution right?
3. The assumption that AI can be a substitute of people (designers and users) in the process
What I was suggesting before was that AI can make us limitless, that it can boost our creative superpowers and make us more efficient. I was not suggesting that it will replace our taste or common sense. It does not have either.
Also - very importantly - it cannot replace users and user feedback. From all of the nonsense I’ve heard on the topic, the thing I found the worst was the concept of testing with synthetic users. I’ll let Pavel Samsonov’s article convince you of this, he can do a much better job at it than me. But in essence, AI by default will give you a pep talk and some points to consider. If you prompt further it will be harsh and more critical. But still, it won’t give you real user insights. As Pavel said, it won’t buy your product. AI agents won’t be your users either, that’s another piece of nonsense.
AI might help you with your output but it won’t help you in defining and validating your problem. Only people (designers and users) can do that.
4. Creative income and livelihood, intellectual property — nothing is off limits.
One more thing — this HUGE, non-ethical elephant in the room… The biggest heist of intellectual property and creative income in history… That’s probably the thing I hate the most about AI. And you might think one AI provider is better than another in this regard but they are all non-ethical to an extent.
I mentioned before that we can all become limitless designers and generalists now. And while that’s good for some of us for now, it’s very bad for creators in general in the long term. You will need creators less, pay them less and eventually you won’t need them and pay them anymore. Instead, you will pay OpenAI, Midjourney, Apple, Microsoft, Nvidia, Meta, Canva, Figma, Adobe, Alphabet.
And you won’t pay junior and non-influencer designers anymore, you will pay a few designers that have mastered the use of AI and can do the work for 10–100 others. Which can mean 9–99 are out of a job. All in all, this is a big power grab and a value grab. That value will flow towards big tech and a few AI powered designers.
Same as you, others can have limitless powers now. And use that power to grab a bigger piece of the pie. Someone that was specialised in something else might step in your area easily. AI Bootcamp will be the next UX Bootcamp creating a new inflow of workers in a already destabilised market.
Love it or hate it?
Or both?
Whatever you feel - its valid and its fine. We’ll be fine. Just like the dog in that meme.
Just another fire we need to tame. Or stay calm while it’s raging.
This article was originally published on Medium. You can find the original post here.