So many of the same, but different colours… Colour tones.

I’m not a painter or any type of artist or graphic professional who pays attention to the colour details or knows colour theory. I know some basic stuff for day-to-day life; white pants are a challenging choice – brave style, hell, to keep clean, dressing all in black is very not cool to some and very stylish to others (funerals excluded), guys don’t buy pink cars, if someone’s face turns blue – better get a doctor on the case asap.

And living my whole life in different big cities – I also absorbed the “city pallets”. From century-old red bricks, to boring grey concrete, to strange soft tones, to bold attempts to stand out.

And when you live in the cities for so long, you also notice that keeping up the colour of any structure seems to be a tough task for the city services.

Well, I get it. Even without any deeper research, the reasons can be relatively obvious.

• Lost original paint (even its formula or ID) and grabbed something “similar.”

• Weather elements make the original colour fade/change, hence even the same paint looks different.

• Dirt on the original layer (not everyone washes the whole wall before painting it over…)

• Can’t be bothered…

And doing a bit of research, the list continues:

• The way you apply the paint can influence the final colour

• Regulatory changes might start dictating what is light-blue in 2010 vs light-blue in 2020 (never trust your eyes, trust the stamped and signed paper!)

• Metamerism

• Intentional temporary fix due to the plans for a full repaint in the near future

Well, the temporary fix falls under one of my most favourite sayings:

There is nothing more permanent than temporary.

But if you walk the streets of any city and pay attention to this, you’ll notice this case a lot.

I kept thinking about how to avoid it, how to fix it.

And some fixes are obvious (not all are simple, though):

• Colour-matching

  • On site, in a different light, a couple of measurements, a couple of tests – can get you the right formula.

• Gradient edges / blend it in

  • Stop with hard edges. Just blend it in. However, there are such massive colour mismatches that it will just look like a massive stain. But in cases with “we are pretty close,” – might work out well.

• Reduce the pallet, keep stock, keep a track, and a knowledge base on how paint fades and which tone to use after 1 year, 5, 10 …

I’m sure the list can go on with much smarter and better solutions. But, there is one solution that rules them all:

Just don’t care!

Cities are… busy. Vibrant. Moving. Changing. Things get built, things break. A patchwork of paint is probably not as critical an issue. Some cities still try, some have given up. Doing all those solutions and more equals time, money, patience, and commitment.

And after all, I was not going to try and solve it here. I was only going to share a small set of photos I took from 100s of patches that I’ve noticed through the years. These are just fun artifacts we get to live among.

• Happy

• Curious

• Worried

First, let’s cover Happy and Worried:)

AI creations are getting marked! You can start noticing this on LinkedIn as they introduced Content Credentials.

It looks something like this:

Then, relatively recently, there was a proposal to add HTTP headers clarifying whether the served content/page contains AI-generated content.

It’s still a proposal, and I’m not sure it’s moving forward much.

We have also lived through a short era of text checkers for AI generation. It was fun, until those checkers confirmed that the Constitution of the USA was written by AI.

We have trained our eyes at this point to spot the AI-generated imagery most of the time, and we sort of can get a sense of whether a chunk of text from someone was generated by AI or written genuinely by them. Our accuracy is not 100% and maybe even not 80%. But we got good at it.

As we try to identify what is genuine human creation and what is AI, we seem to still be confused about what to do in the cases of AI generation.

On one hand, we sell multiple solutions that are based on AI generation. All shapes and forms. Mixed media. Endless use cases.

On the other hand, we also hunt down with pitchforks those who use AI-generated content in some industries. I.e., Clair Obscur case.

On Hacker News, there is a comment from the user nrmitchi with a good point. Here is a piece of the comment and a link to the full one:

Yes, there is a huge problem with AI content flooding the field, and being able to identify/exclude it would be nice (for a variety of purposes)

However, the issue isn’t that content was “AI generated”; as long as the content is correct, and is what the user was looking for, they don’t really care.

This is a cornerstone of the confusion, as it seems to me. AI is still a probabilistic technology with its precision getting better and better; however, knowing that it’s not deterministic, we tend not to give it our full trust.

All above explains my happiness (being able to have a regulatory way of marking human vs. AI) and worry (aren’t we over-regulating? If it’s valuable, then who cares how it was created?).

And now to the case that I’m curious about. It’s the case that’s not yet TRUE. It’s not yet implemented or enforced. But as I observe the developments, emotions, EU’s love for regulations (moats…) – I can see the scenario where marking the AI-generated content becomes a regulation/law.

And in that scenario, I’m wondering what it would do to the fashion industry, specifically the e-com part. How would that impact the decision-making of a common consumer?

If today we shop like this:

Review items like this:

But then, with the marking in place, if we get to shop with something like this (AI mark mock-up):

And like this:

What would this do to the consumer choices, trust, and preference? I’m very curious to see the A/B test on items like this. As nrmitchi said above – if the result is correct, no one would care if it’s AI or not. But, how would our consumer psychology work if we are looking at something that maybe is correct, however, it’s still MARKED as AI. Wouldn’t we think twice? Thinking “Hmm..they mark it as AI because it might deviate from how it actually is?”

It could also take one friend or friend of a friend who “bought an AI-marked item, and it was really off – I don’t trust these images anymore” – however, in reality, it could still be the usual e-com consumer experience, as even photos of the real items don’t always convey how the items will really look ON YOU.

If I were to make predictions and place bets, I think that if we get to the state where marking AI-generated content becomes a must, fashion would need to rethink its usage in direct-to-consumer content. I would imagine that an online shop mixed with real shots and AI shots wouldn’t be the best of ideas, and brands would have to either go all AI and build the trust with the consumer that their AI shots are CORRECT, or keep it “old skool” and keep the craftsmanship of the photoshoot alive.

I do want to witness that “conundrum” even though it might cause some negative impact to my professional plans… But I think it’s an interesting experience that we will all learn from. So… Bring on the markings!!!