Tag Archives: Stable Diffusion

My First Attempt At Creating AI Art With Stable Diffusion

I cannot draw worth a damn.

In fact, that’s an understatement. I would go so far as to say I’m so terrible at drawing that most 10-year-olds couple probably match or exceed my drawing skills.

Believe me, I’ve tried to draw things. My love of comic books inspired me early on to try. But I learned very quickly that this is just not a skill that I have, nor is it a skill I can ever be good at. No amount of classes can make me good at it. No amount of lessons, encouragement, or practice could ever make me halfway decent at it. I genuinely wish I had some drawing skills, but I don’t.

Now, I’m going out of my way to bemoan my terrible drawing skills for a reason. For much of my life, I’ve accepted that this just isn’t something I can do. Everybody has shortcomings. This just happens to be mine. And for the most part, I’ve been content to live my life knowing I’ll never draw anything worthwhile.

Then, AI-generated art started to emerge. And suddenly, there’s an alternative. It may just give people like me hope that they can one day create genuinely beautiful artwork without any tangible drawing skills.

I mentioned it recently, but I don’t think I adequately conveyed just how exciting this new technology is for someone like me. I’ve talked a lot about the potential of artificial intelligence, good and bad. But a lot of that was just me speculating on the potential. This is a real, usable product of artificial intelligence that anyone with an internet connection and basic language skills can use right now for free.

This isn’t some fanciful tech from a fictional future. This is real.

It might not be overly advanced in that it can perfectly turn out thoughts into an image. It’s still relatively crude in that it basically just takes images from the internet and uses them to mix, mash, and compile images based on prompts. But like all emerging technology, it has to start off crude. Over time, people and organizations implement refinements. The product gets better, more advanced, and more efficient.

Eventually, it becomes so capable that we marvel at how crude it used to be. Just show anyone under the age of 20 an old flip-phone for proof of that.

For me personally, AI-generated art is more than just a toy or a novelty. It’s a way for me, a man with no ability to draw anything, to turn ideas into an image. I can’t overstate how powerful that is for some people. And I really do look forward to seeing this technology grow in terms of capability and efficiency.

To demonstrate just how remarkable it is, I used an AI-generating art programs on the web called Stable Diffusion to create some art. This is something anyone can use right now for free. Granted, this is a beta version so the results aren’t going to look overly professional. But for something that’s free and easy to use, it’s still remarkable.

Below are just some of the images I created, along with the prompts I used.


A beautiful woman admiring a sunset.


A Christmas tree in the forest during a snowstorm.


A lush Amazon jungle with a river.


A futuristic city skyline.


A female android with blue eyes.


A shadowy figure in a forest.


A couple walking through a forest during a snowstorm.


A man running from a forest fire.


A woman standing at the summit of a mountain.


A tornado forming in a city.


Again, I cannot draw. I cannot paint or create images on a piece of paper or a computer screen. But thanks to Stable Diffusion, I was able to create these in just a few minutes. It cost me nothing and it was a genuinely profound experience, turning ideas into an image.

And keep in mind, this was just the free beta of the program. There are more advanced programs you can test out, but you do have to sign up and apply for their use. There are also some versions that you have to pay for. Once this technology gets especially refined, that might be something worth the money. Knowing I’ll never be able to physically draw, it’s definitely something I’ll consider.

For now, I’ll be keeping a close eye on this emerging technology. If I create more quality AI-Generated art, I’ll be sure to share it too.

And if you’re like me in that you don’t have good drawing skills, or any drawing skills for that matter, check out Stable Diffusion. Create some art that you never would’ve been able to otherwise create. It’s a great feeling and something that might help get you excited about the future of creativity.

1 Comment

Filed under AI Art, Artificial Intelligence, technology

The Exciting/Distressing World Of AI-Generated Art

Whenever I talk about artificial intelligence, I often talk about the possibilities and potential it has for the near and distant future. I admit I sometimes to a bit overboard with the speculation and the hyperbole. In case I haven’t made it abundantly clear already, I am not an expert. I do not consider myself exceptionally smart or well-informed on this topic. I just find it very interesting and quite tantalizing, given how much I’ve seen computer technology evolve over the course of my lifetime.

However, in talking about artificial intelligence, I rarely get a chance to talk about some actual tools and products powered by AI that we can use today. That’s just the nature of technology like this. It takes a while to develop and refine. It also takes a while to actually create a usable product with it that don’t require a Masters Degree in computer programming.

But this past year has seen the rise of a new type of AI-powered product that is making its presence felt. It’s called AI-Generated art and it’s exactly what it sounds like. It’s art entirely created by an artificial intelligence that uses massive amounts of data on art, shapes, and design to craft artwork based solely on text suggestions.

It’s not entirely new. For a number of years now, there have been AI systems that can essentially create photo-realistic depictions of people who don’t exist. That, alone, is an impressive feat and one that has some distressing implications for those worried about fake IDs, identity theft, or catfishing. However, these new AI-Generated art programs have the potential to do so much more.

While the mechanisms behind it are very complex, the interface itself is very simple. A user just enters a brief description of what kind of artwork they want. Then, the program processes that and crunches the data. Finally, it generates an image. Sometimes, it takes a few seconds. Sometimes, it takes a bit longer, especially if the prompt is more elaborate. If you want to see a good example of what it can create, just check out the brief, but hilarious skit John Oliver did. Just don’t watch it while eating cabbage.

Aside from the inherent comedy gold that can be mined from this technology, just take a step back and consider the larger implications of these tools. These are prompts being entered by people who probably don’t have much in terms of art skills. And as John Oliver noted, some are being entered by people who aren’t entirely sober. While the images they generate don’t exactly look like masterpieces or anything someone would mistake for photo-realism, it’s still remarkable they’re as good as they are.

In seeing some of this art, it actually reminds me somewhat of early video game consoles that began rendering 3D graphics. I’m old enough to remember the somewhat clunky transition between 2D to 3D graphics. Just look at early Playstation games or games like Super Mario 64. They weren’t exactly polished, but they were a step in that direction.

Now, compare that to a typical game on the Playstation 5. In the span of just 20 years, the graphics and renderings have become so realistic that they’re navigating uncanny valley territory. With that in mind, imagine what these AI-generating art programs will do with that kind of refinement. I don’t know if it’ll take 20 years or longer, but it does create some tantalizing possibilities.

Ordinary people could conjure detailed, photo-realistic backgrounds for games, portraits, or stock art.

Ordinary people could conjure elaborate scenes and illustrations for stories they wish to tell.

Ordinary people could create artistic depictions of elaborate fantasies, including the sexy kind.

This is especially intriguing for someone like me because, as I’ve noted in the past, I cannot draw worth a damn. I have practically no skills when it comes to creating visual artwork, be it with a pencil or a computer program. I’ve never had that skill. I’ve tried many times in the past do develop those skills. I’ve never succeeded. I’ve always been better with writing and words. And I’ve been perfectly content with that.

Now, this technology gives someone like me an opportunity to craft images to go along with my words. It opens the possibility that I could one day write a story, sexy or non-sexy, and supplement it with real, vivid depictions of the characters and scenes. That is definitely something I want to pursue. I have experimented a bit with the AI art programs, but they’re still someone limited. I won’t be incorporating them into my sexy short stories anytime soon.

But if these programs continue to improve, then it’s only a matter of time before I craft a story in that manner. Honestly, that really does excite me, more so than a lot of the promising news surrounding artificial intelligence. I understand there are aspects to the technology that may never happen or just won’t be happening within my lifetime. But these AI art programs are real. They exist now and they’re going to be refined, like most emerging technology. It remains to be seen how fast they’ll achieve a higher quality, but I will certainly be watching it closely.

If you want, you can even test these programs out yourself. This site lists 10 sites you can go to right now, but these are the sites I recommend.

Dall-E 2

Artbreeder

DeepAI

StarryAI

Please note that most of these services are limited and none allow anyone to create images that are overly pornographic or outright illegal. However, you can still create some legitimately good images, which you can save and use in whatever way you please. I’m already hoping to use some for my YouTube channel.

But even though this technology is especially intriguing to people like me with no art skills, I don’t deny it has actual artists very concerned. There has already been one instance where an AI-generated artwork won an art contest, which the real artists did not appreciate. It’s not just that an AI like this won without putting in the effort an artist usually would. In many cases, these programs used art other artists had created to refine its code. Over time, these programs could conceivably put those same artists out of work.

I can totally understand that concern. Who would hire a talented, but expensive artists to create images if they could just use an AI program to create it in seconds and for free? Do you really think big companies like Disney, Warner Brothers, and Universal wouldn’t fire their entire art team if they could get the same results for a fraction of the cost? They’re billion-dollar profit-driven companies. You know they would.

Even if this technology doesn’t completely replace real-life artists, it’s still essentially doing most of the work. On some level, it dehumanizes the artistic process, even more so than a camera. A camera can only render the image in front of it. These programs could conceivably conjure images that nobody has ever seen or imagined, a feat that once belonged solely to artists.

What does that mean for the future of artists?

What does that mean for the future of art?

I don’t claim to know the answers. I’m not even sure how to speculate on something like this. Again, I have no art skills with respect to drawing or creating images from scratch. I’m the kind of person who will embrace this technology more than most, so I’m going to be somewhat bias in that regard.

But artists and governments are starting to take notice. China has already made waves by attempting to ban AI-generated media that isn’t appropriately marked. While that may temper some trends in this field, it’s not going to stop it. There’s just too much to be gained at this point. The genie is out of the bottle and there’s no putting it back.

It’s sure to cause more issues, especially as the technology becomes more refined. It probably won’t be long before a major problem occurs because someone used AI-generated art in some nefarious way. Some are already trying, but they can only achieve so much, given the limits of technology.

That will eventually change. If you’re reading this, you’re likely to see some AI-generated artwork that you’ll mistake for something real. At that point, even concerns about deep fakes will be minor in comparison. Only time will tell.

Until then, non-artistically inclined people like me can start contemplating what thoughts and ideas we can one day make real.

Leave a comment

Filed under Artificial Intelligence, futurism, technology