Design employers want Adobe Illustrator, InDesign AND Canva/chatgpt, experience!

[WildWestDesigns]

You're all right about the quality of graphic design being hit hard with this technology, that "isn't there" yet. Sadly clients don't know, don't see and worse, don't care about the difference. They can't pay for the service anymore if they want to stay afloat.

Ironically, this isn't the first time. This always happens when newer tech comes about. This just feels like it's on steroids though. Most normies, no matter what area one is talking about, don't really care and when they don't care about differences, they go with price, because that affects them in other ways moreso compared to those "differences". Output that is related to production of some type is probably what's going to be safe, unless it's style of "learning" changes as there is some much "stuff" that they learned that is just bad for production. Downside is if people stop going for physically produced items. I don't know if that will totally be the case, people in the end still like something that they can touch, so I don't know about that. The question becomes how this tech can handle final production and how much hand holding it does or doesn't need to accomplish that task.

 

[Johnny Best]

Thinking 'bout the time you drove in my car. Thinking that I might have drove you too far. And I'm thinking about the love that you laid on my table." (Cream/Badge).

Cream put out some great sounds!

 

[Bobby H]

If many of you think that AI isn't a menace, you're really shoving your head in the sand. AI is progressing like crazy. If you're using free versions, yes, you can believe that AI isn't there at all. Those free solutions are limited. As a sign designer, my job isn't quite erased from earth, but it's very damn well hit by it.

In my experience so far, artwork from so-called AI has created more problems than it has solved. Yeah, small business customers are increasing their use of AI-based tools to develop things like their logos or other creative elements. But then somebody like me has to come in and fix a lot of crap in the AI-generated artwork and then get it re-drawn in vector-based form.

There is a lot more to sign design than just coming up with a logo or other original artwork.

I have yet to see an AI-based tool create sign designs that are accurate to scale. Properly designed signs have to be tailored to a given space. That means creating designs for replacement faces going into existing cabinets, channel letter signs that fit a certain foot print on a building fascia and on and on from there. None of the AI-generated artwork is production-ready. The AI bots cannot seem to create clean, vector-based artwork. The bots certainly do not know how to automatically generate accurate shop drawings with dimensions, materials and spec lists, etc.

Here's a bigger problem with the AI stuff: it really shits the bed when it tries to do anything involving larger companies with established brand identities. The AI bots can rarely ever generate the same exact image twice. That's why I say I feel like I'm playing a slot machine when I generate images with those tools. I don't know what the hell I'm going to get. An established national brand will have specific colors, white space rules and other handling requirements. The AI bots don't know how to handle that stuff since they are abstract things.

As I've said before, the AI tools can be useful for brainstorming artistic concepts. They're not much good for anything else beyond that.

 

[TetroDesign]

I have yet to see an AI-based tool create sign designs that are accurate to scale. Properly designed signs have to be tailored to a given space.

You should check out Chaos Veras. An AI application, extension of Nano Banana. Professional AI stuff.

Check out the image I gave it. I started out drawing a flat design. It automatically detected everything!

 

[netsol]

In my experience so far, artwork from so-called AI has created more problems than it has solved. Yeah, small business customers are increasing their use of AI-based tools to develop things like their logos or other creative elements. But then somebody like me has to come in and fix a lot of crap in the AI-generated artwork and then get it re-drawn in vector-based form.

There is a lot more to sign design than just coming up with a logo or other original artwork.

I have yet to see an AI-based tool create sign designs that are accurate to scale. Properly designed signs have to be tailored to a given space. That means creating designs for replacement faces going into existing cabinets, channel letter signs that fit a certain foot print on a building fascia and on and on from there. None of the AI-generated artwork is production-ready. The AI bots cannot seem to create clean, vector-based artwork. The bots certainly do not know how to automatically generate accurate shop drawings with dimensions, materials and spec lists, etc.

Here's a bigger problem with the AI stuff: it really shits the bed when it tries to do anything involving larger companies with established brand identities. The AI bots can rarely ever generate the same exact image twice. That's why I say I feel like I'm playing a slot machine when I generate images with those tools. I don't know what the hell I'm going to get. An established national brand will have specific colors, white space rules and other handling requirements. The AI bots don't know how to handle that stuff since they are abstract things.

As I've said before, the AI tools can be useful for brainstorming artistic concepts. They're not much good for anything else beyond that.

a friend is operations manager of a good client. he is automating repetitive tasks.
his problem is if the result is acceptable the first time he tries, it is rarely repeatable
no one goes through the multiple extra steps "creating a project, then a version, each time a minor change is made
you see the perfect result, but, may never get back to it again

i took a "dashboard" output in excel, that was exactly what they wanted., but, not repeatable
i cheated and used the ms access developer kit to be able to "install and run it as an app"

he laughed and said i showed my age "no one does it that way, they end of lifed the developer kit 10 years ago"
but, MY WAY WORKED

 

[Hal D]

a friend is operations manager of a good client. he is automating repetitive tasks.
his problem is if the result is acceptable the first time he tries, it is rarely repeatable
no one goes through the multiple extra steps "creating a project, then a version, each time a minor change is made
you see the perfect result, but, may never get back to it again

i took a "dashboard" output in excel, that was exactly what they wanted., but, not repeatable
i cheated and used the ms access developer kit to be able to "install and run it as an app"

he laughed and said i showed my age "no one does it that way, they end of lifed the developer kit 10 years ago"
but, MY WAY WORKED

Right on, my brother!

 

[Hal D]

I tell prospective clients up front that if you're looking for a "graphics designer", then you're in the wrong place. I specialize in signage, architectural elements and environmental graphics. Nothing more and nothing less.

 

[netsol]

To me using AI (whether to automate tasks or to prepare artwork) is like buying a lottery ticket.
you COULD get a winner, but can’t DEPEND on it

well over 1-1/2 years ago, another client’s HR director hired an AI DEVELOPER

he seems to be a success. The only thing he does is keep employees working on these projects focused on repeatability & pulling the plug EARLY on efforts that are not destined to succeed.

 

[WildWestDesigns]

To me using AI (whether to automate tasks or to prepare artwork) is like buying a lottery ticket.
you COULD get a winner, but can’t DEPEND on it

Unfortunately though, most are depending on it. Especially if they can't go any further in the endeavour that they are in.

I just wish people would stop using the term "AI". What passes for "AI" really isn't "AI". Not at this point. It is effectively power/live trace 2.0 and/or LSP/auto complete 2.0 (depending on it's application) and that still has narrow areas to where it's going to be successful. I would argue that if a person knows what they are doing, the amount of time/effort that it takes to continually prompt (which is going to have to be another layer of abstraction given the poor readers that are coming out of school nowadays at least here stateside) and clean up what is rendered back takes more time compared to just doing it the good old fashioned way from the get go. But the key thing is that the person knows what they are doing.

Now, I'm not against automation, I wish people were more aware of what is being automated away, so that way when things happen, work doesn't come to a stand still when their "AI" service has gone done for whatever the reason (or if their script has to be tweaked due to a change in the program's API).