Jump to content
You must now use your email address to sign in [click for more info] ×

Recommended Posts

8 hours ago, R C-R said:

Vague, cagey comments about future software developments like this are not unusual in the industry

True, but not so much when you are coming from behind to just play catchup and there is nothing that anyone else could gain or copy from your statement as it puts no new ideas out into competition. Affinity just revealed they are working on scripting/plugins and IMO this is actually more significant to competitors as it will allow Affinity to more rapidly close the gaps. At this stage, the default assumption in the industry is that everyone is scrambling to implement AI. It would not be a surprise or even useful to competitors.

Link to comment
Share on other sites

When it comes to customers AI fans think that it is lost opportunity if you dont use AI. Some threatening that they will leave to competition. But reality is that there is equally big group of users who is in principal against AI. And they are equally valid customers that will pay for tool without generative AI tools. The moment you leave Affinity for Adobe because they have AI - somebody from Adobe will come to Affinity because it doesnt have it :D Simple as that. Frankly I dont know if its not good marketing strategy for Affinity to say "NO SUBSCRIPTION, NO AI, NO PROBLEM!" hahaha

Few months ago Clip Studio Paint wanted to add image generation tool - it was met with huge backlash from their user base and they rolled back those plans. Artstation didnt wanted to ban AI on their site and immediately we got great new portfolio sites that are not accepting AI like artgram or cara.

Those two groups - pro and against AI are equally paying customers. You cant please both. Either you add AI tools or you dont. One of the groups will not like it. So really Affinity by not having these tools now is not loosing any more users then it would if they added them.

Link to comment
Share on other sites

14 minutes ago, nezumi said:

But true is that there is equally big group of users who is in principal against AI.

Show me the number of podcasters who just made videos saying they are leaving Adobe because they are adding AI. I would be very interested in watching them.

Doesn't exist in the way you describe it and certainly not equal. I state that as someone who has been more critical than most on the topic of generative AI.

Here is a sample piece I created from my own perspective from an article I wrote about AI.

image.png.9a2eef4dfc897d82c7e6f8732429dce3.png

AI is far more than just the random image generators like Midjourney. It includes image scaling, object removal, auto selections, filtering, noise removal and the list continues to grow. It is also evolving into more useful tool like applications like Adobe's content fill.

Nonetheless, given Affinity's resource constraints, my assumption is it would likely be best for a number of reason for Affinity to simply ensure we have the ability of doing our own integrations. AI is evolving so fast that anything integrated will be obsolete nearly as soon as it is shipped.

Link to comment
Share on other sites

Hahaha, why podcasters? Is that how you look at the world? Through YT videos? I just gave you examples of events that happened - entire portfolio sites filled with people that is against "AI art" for one reason or another. But lets ignore it all because podcasters :D

And sorry but you have not created anything. You wrote what you want to have created FOR YOU. Plus added specific artist name.. Quite frankly that is sad more then anything else.

Link to comment
Share on other sites

5 minutes ago, nezumi said:

Hahaha, why podcasters? Is that how you look at the world? Through YT videos? I just gave you examples of events that happened - entire portfolio sites filled with people that is against "AI art" for one reason or another. But lets ignore it all because podcasters :D

And sorry but you have not created anything. You wrote what you want to have created FOR YOU. Plus added specific artist name.. Quite frankly that is sad more then anything else.

Talk about totally missing the point. Wow. Did you even read it?

Link to comment
Share on other sites

As so often here in the forum, these are endless and pointless pros and cons debates on the subject of AI. Because there will always be people who are for or against it, or who think differently than others about that theme (since people do have different opinions here). - That's why we're just always only going in circles here for such threads.

☛ Affinity Designer 1.10.8 ◆ Affinity Photo 1.10.8 ◆ Affinity Publisher 1.10.8 ◆ OSX El Capitan
☛ Affinity V2.3 apps ◆ MacOS Sonoma 14.2 ◆ iPad OS 17.2

Link to comment
Share on other sites

@CM0

I apologize. Frankly you almost lost me at "show me podcasters" and I gave up on "piece I created" presenting generated image. Sorry I have seen too much of it to keep going first time around. Guilty as charged.

Now I read it.

In other part of this forum I have spoken about AI tool for tracing bitmaps. Works very well - most accurate tool for tracing I have used. I am using Davinci Resolve also with quite good AI tools. As a 3D artist I am of course using denoisers to speed up rendering. These are actually helpful tools not just "I will do that for you" image spitters and hell, I will happily welcome that sort of AI in Affinity too. I think sooner or later it will become norm. So I am 100% in agreement with what you said. I just really dislike generative AI spitting ready artwork. Its just... wrong on many levels and sad on many more.

@v_kyrpart of the fun. Obviously nobody will change anything by posting on forum his opinion. Doesn't mean we cant have meaningless yet entertaining conversation for sheer fun of it.

Link to comment
Share on other sites

4 minutes ago, nezumi said:

I just really dislike generative AI spitting ready artwork. Its just... wrong on many levels and sad on many more.

Thanks for taking a moment to have a closer look :)

Yes, I totally agree with that point of view.

Link to comment
Share on other sites

2 hours ago, nezumi said:

Doesn't mean we cant have meaningless yet entertaining conversation for sheer fun of it.

Always thought that's only the special domain of BratGPT. 🤔

A small Q/A conversation with Brat ...

brat1.jpg.23d3d182a6840f372ef757177275b1d5.jpg

brat2.jpg.f93936504f0f6570c0bafc2429942239.jpg

brat3.jpg.1e973f7748ae27188a9d2e05d3e58d3b.jpg

brat5.jpg.42fd1ab90c78d9823637f076496e9b19.jpg

brat6.jpg.ba70efa4be25d0ea6878dc91258d4f86.jpg

So much for that! 😉

☛ Affinity Designer 1.10.8 ◆ Affinity Photo 1.10.8 ◆ Affinity Publisher 1.10.8 ◆ OSX El Capitan
☛ Affinity V2.3 apps ◆ MacOS Sonoma 14.2 ◆ iPad OS 17.2

Link to comment
Share on other sites

I believe Generative fill is just impossible with Affinity.
It would be challenging to handle the costs associated with cloud processing and training photo dataset licenses.
Photoshop can do that because its subscription-based, has huge cloud infrastructure and own Adobe stocks.

It would be more feasible to integrate Affinity with a third-party AI generative images provider, which allow users to have the choice to subscribe to such services if they want.

So Affinity can focus on their thing.

I would still be excited to see basic AI tools integrated into Affinity, such as object selection, sharpening, enlargement, etc
those are useful for photo editing and does not rely on cloud processing.

Link to comment
Share on other sites

53 minutes ago, bayustudio said:

It would be challenging to handle the costs associated with cloud processing and training photo dataset licenses.

I agree. That is the most logical assessment for Affinity not making a direct commitment.

54 minutes ago, bayustudio said:

I would still be excited to see basic AI tools integrated into Affinity, such as object selection, sharpening, enlargement, etc
those are useful for photo editing and does not rely on cloud processing.

Indeed, there are still other opportunities to leverage the technology. Furthermore, I still think finishing their APIs for plugins/scripting may be the fastest path for us to benefit from some of these capabilities as the community and 3rd parties can start making some of these capabilities available via integrations.

Link to comment
Share on other sites

On 5/26/2023 at 2:48 AM, WMax70 said:

90% of what you wrote I agree with.

However, about the ship service engineer looking at the vessel you also said:
"It looks exactly like quickly slapped person from another photo with little to none effort to match lighting..."

Let's analyse the light source in this image.
It's mainly the sun, shining from the rear towards us. So it  should light the front of the person, not the rear.
I would even suggest the persons back might even be more dark than in originally image, but for sure no more lighted like you did.
The sun light can touch the right schoulder cause it is reflective material but not the back of the person. 
Therefore I disagree with your changes. Now it does not match anymore with the light source. 
Imo, you inserted false or at least a non natural light source to lighten the back of the person. However, this might go of topic.
 

Nezumi's quick editing did fix the colours of the guy so he fitted better into the background. The colours are too blue and cold for such a warm and gold coloured scene. Some of the results from the AI just do not match the the parts that are retained from the original images. There will still be a need for manual editing so the results do not look like a bunch of different images thrown together. The AI is not smart enough yet to do all the work for you.

Link to comment
Share on other sites

The biggest hurdle that AI will have is the legality around using it for commercial work. The tool in Photoshop can't even be used for commercial products according to Adobe, so it is going to be a legal minefield when using it in any professional capacity. Also, the amount of rules surrounding this tool are very restrictive. While I agree with a few of them, the rest are going to really prevent it from being widely used in certain fields (also not a fan of how much surveillance that Adobe will have on their users through these AI tools).

If there are any AI capabilities I would like to see in Affinity that aren't image from prompt, it would be what @bayustudio mentioned like AI upscaling, sharpening, etc. Those to me are way more useful and more feasible for Affinity to implement than trying to pull off what Adobe are making with their vast content library. There are already some open source alternatives for these types of features that are more than capable, so it would be highly appreciated if they got added to Affinity as well. AI upscaling especially would be great for vectors once Serif decide to implement raster-to-vector into Designer, since larger images tend to produce better vector results.

Link to comment
Share on other sites

Maybe we’ll get generative fill through a third party plug-in.

There already are some AI plugins that work with Affinity Photo.

Advertising designer - Austria —  Photo - Publisher - Designer — CS6 d&wP — Mac Pro 5,1 (4,1 2009) 48GB 2x X5690 - RX580 - 970EVO - OS X 10.14.6 - NEC2690wuxi2 - CD20"—  iPad Pro 12.9" gen1 128 GB - Pencil

Link to comment
Share on other sites

13 minutes ago, BofG said:

Do you have a source for this?

From what I have found, Adobe states they trained the AI on their own stock libraries and the generative fill at least is safe for commercial use:

https://www.adobe.com/uk/products/photoshop/generative-fill.html#:~:text=Push creative boundaries with more peace of mind.,is designed to be safe for commercial use.

Here are the rules for using the AI tools from Adobe as of now. They are strictly prohibiting commercial use. Also, there was a ruling in America some time ago that judged that AI generated imagery cannot be copyrighted, only the parts that you yourself own and/or create by your own hands can actually be copyrighted. If you are planning on using AI you will have a hard time protecting your work through copyright since making prompts are not going to be enforceable unless something drastic changes with copyright and ownership laws. There are plenty of companies that refuse to use AI for this and many other reasons.

Here's another video discussing the ruling if you want to learn more.

 

Link to comment
Share on other sites

9 minutes ago, BofG said:

From what I have found, Adobe states they trained the AI on their own stock libraries and the generative fill at least is safe for commercial use:

I believe the limitation is applied only for beta.

Link to comment
Share on other sites

1 minute ago, Frozen Death Knight said:

If you are planning on using AI you will have a hard time protecting your work through copyright since making prompts are not going to be enforceable unless something drastic changes with copyright and ownership laws.

It is actually rather easy. You only need to modify the image and you have human created content. The copyright ruling is going to be practically unenforceable except in exact duplication from some of the generating websites.

Link to comment
Share on other sites

6 minutes ago, CM0 said:

It is actually rather easy. You only need to modify the image and you have human created content. The copyright ruling is going to be practically unenforceable except in exact duplication from some of the generating websites.

The way that the Adobe tool does it makes everything it generates not copyrightable according to said ruling. Unless you heavily modify those areas afterward you will be unable to enforce copyright, which most of these examples shown by YouTubers aren't doing. Slapping a few filters on are really not going to be enough.

Link to comment
Share on other sites

13 minutes ago, BofG said:

You said "according to Adobe". Where is the official information from Adobe?

The guy is reading directly from Adobe's own website in that video (notice "adobe.com" at the very top):

image.thumb.png.215437f0d46422ef9dbd6747b924cb9d.png

 

image.thumb.png.01ddf49534eb6d85e05acb82124c0bc0.png 

Even if Adobe "allow" commercial use in the official release, they do not have the authority to determine its viability as a commercial product, the law is. The law has not yet been written and ironed out in regards to AI, but there has been legal precedent set by a court ruling that determined that AI generated imagery is not copyrightable, as previously mentioned. We won't know how AI will be used for commercial purposes until more laws have been written. 

Link to comment
Share on other sites

41 minutes ago, Frozen Death Knight said:

The way that the Adobe tool does it makes everything it generates not copyrightable according to said ruling. Unless you heavily modify those areas afterward you will be unable to enforce copyright, which most of these examples shown by YouTubers aren't doing. Slapping a few filters on are really not going to be enough.

There have been recent changes to the copyright position allowing for mix of AI and human. Furthermore, it will be unenforceable because ultimately nobody will ever know as it continues along the path of inevitably becoming indistinguishable.

https://www.theartnewspaper.com/2023/05/04/us-copyright-office-artificial-intelligence-art-regulation

Link to comment
Share on other sites

Just now, CM0 said:

Furthermore, it will be unenforceable because ultimately nobody will ever know as it continues along the path of inevitably becoming indistinguishable.

Unless AI's can be trained to tell the difference? 🤔

All 3 1.10.8, & all 3 V2.4.1 Mac apps; 2020 iMac 27"; 3.8GHz i7, Radeon Pro 5700, 32GB RAM; macOS 10.15.7
Affinity Photo 
1.10.8; Affinity Designer 1.108; & all 3 V2 apps for iPad; 6th Generation iPad 32 GB; Apple Pencil; iPadOS 15.7

Link to comment
Share on other sites

Just now, R C-R said:

Unless AI's can be trained to tell the difference? 🤔

This is already been tried, but it is not accurate. Not because it can not do so, but because the evolution continues changing. Furthermore as it continues to get better and better you end up with false positives.

Link to comment
Share on other sites

9 minutes ago, CM0 said:

There have been recent changes to the copyright position allowing for mix of AI and human. Furthermore, it will be unenforceable because ultimately nobody will ever know as it continues along the path of inevitably becoming indistinguishable.

https://www.theartnewspaper.com/2023/05/04/us-copyright-office-artificial-intelligence-art-regulation

image.png.8df13e57e843bf739e0416073660da2e.png
image.png.1a5ca54e245c900cdc510cd576037c1e.png
The two most significant quotes from the article that reinforce what I mentioned before. AI art is not copyrightable. However, with enough changes a human may claim authorship, but that is not yet clear. In the case of the book mentioned in that article, it means that every single image on it and inside it can be used by anyone without infringing on any copyright. The latter quote is still up in the air in regards to how transformative the work is to be claimed authorship over. As the article mentioned, it's going to be judged on a "case-by-case basis".

Link to comment
Share on other sites

4 minutes ago, Frozen Death Knight said:

The two most significant quotes from the article that reinforce what I mentioned before. AI art is not copyrightable. However, with enough changes a human may claim authorship, but that is not yet clear. In the case of the book mentioned in that article, it means that every single image on it and inside it can be used by anyone without infringing on any copyright. The latter quote is still up in the air in regards to how transformative the work is to be claimed authorship over. As the article mentioned, it's going to be judged on a "case-by-case basis".

Yes, that is what I meant by a mix of AI and human. Nonetheless, it will still be unenforceable as nothing can be enforced that is not knowable. As AI gets better, it will make all IP protections irrelevant, not for just art, but for everything. As Google recently stated, there are no moats with AI.

Furthermore, you don't need copyright for commercial use. The major use cases for this is going to be in social media marketing and social media engagement vs directly selling art. In those instances it doesn't matter.

Link to comment
Share on other sites

1 minute ago, BofG said:

Surely no one thinks that they've spent all this money and time on developing this just for it to be used only by people playing around for personal use.

I've also heard that Adobe plans on using their AI to compensate the original works as their AI can determine which pieces contributed to the generation.

Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

Loading...
×
×
  • Create New...

Important Information

Terms of Use | Privacy Policy | Guidelines | We have placed cookies on your device to help make this website better. You can adjust your cookie settings, otherwise we'll assume you're okay to continue.