(a) change your mindset re: "doomed". No one is "doomed".
(b) Things change. Many people are lazy. For some - "typing" is "work". People like instant gratification. Like SUPER instant. If they have to edit out artifacts, or 'click 3 times' to get an image they like - "ugh" - that is too much work for them.
(c) Sales trends change, sales go up, sales go down, then products are re-purposed and sales go up again.

There are people that like to build/work on their own cars. Then there are people who just like to buy a finished product. Same will be for images & videos.

Your time is never "wasted" - you went in eyes open knowing what the "ai" market was like, and wanted to catch a wave. You still could. Or maybe not. Depends on a whole lot of things.

It is certainly a bit scary, there are people who likely never picked up a camera uploading AI imagery, and there will be people who will automate the process, probably happening already. This same level of concern should also be hitting the agencies who don't allow uploading of AI imagery, I suspect that will change as well....at some point all agencies will accept it.

I can see it happening. I am down to 30-40% which is concerning.
Non-qualified people also have entered the market and have flooded everything with AI.

AI image/video generation should be for licensed professionals only!!   

This is a slow time in the year. No reason to panic.

Many customers are still deciding o their projects for 2024 and this Monday was a holiday i the US, so people take time off.

If everybody is reporting an abrupt  slowdown in sales it is not because of more competing files. those changes come slowly.

Check the search positios of your content. If they haven't changed and you cannot see new better files right next to yours, the slowdown is general sales.

One question I have is that why would people sell AI images on stock agencies because if I understand correctly AI images can not be copyrighted?

No, I think you misunderstood something. There was just a ruling in the US where the copyright office denied copyright to some AI created comic. But the US is not the world - despite what some people think - it's just like 4% of the world population and whatever a judge decides in the US, has no consequences to the rest of the world.
Other countries made different decisions or no decisions about the issue at all.

https://www.technollama.co.uk/chinese-court-declares-that-ai-generated-image-has-copyright

In the United States, a federal judge ruled in 2023 that AI artwork cannot meet federal copyright standards because �Copyright law is �limited to the original intellectual conceptions of the author�.� With no author, there is no copyright.

What am I missing - no human author - no copyright?

I don't understand.

Thanks Derby I'm not in US so I don't know how copyright laws are changed and made law (not asking you to explain) more research on my part.

Thanks again

The easiest tool for copying is not ai but a normal camera.

In fact it is usually a lot faster to copy a bestselling image by just taking a similar picture.

Also customers and clients, if they really want to save money, they can just take pictures with their iphones. Including things like wood backgrounds, green grass and sky etc...a lot of the content that is ultrageneric is very, very easy to take yourself.

And yet here we are.

And if it is true that Shutterstock is getting 50% less content even if they don't take any ai but on Adobe sales are increasing although or probably because they have an additional ai collection, then ai content is our friend that brings more subscribers to the agency that pays us more and treats us well.

The easiest way to deal with ai worries is to just try it yourself. Even just for fun, you don't have to sell it.

It becomes easier to understand that it is just a tool, like Photoshop is a tool.

The easiest tool for copying is not ai but a normal camera.

In fact it is usually a lot faster to copy a bestselling image by just taking a similar picture.

It really depends on what is depicted in the bestseller. It is not so easy to copy a picture of a tropical beach with palm trees or of an iceberg, unless you somewhere near them or travel there anyway.

Also customers and clients, if they really want to save money, they can just take pictures with their iphones. Including things like wood backgrounds, green grass and sky etc...a lot of the content that is ultrageneric is very, very easy to take yourself.

Even images of stuff that is seemingly available everywhere are no always that easy to copy.

For example really clear blue sky with a nice cloud distribution does not occur as often as you might think. And if you want a specific type of clouds like cirrus clouds, you may have to wait a long time until you get an opportunity to get a really nice picture.

The easiest tool for copying is not ai but a normal camera.

In fact it is usually a lot faster to copy a bestselling image by just taking a similar picture.

Also customers and clients, if they really want to save money, they can just take pictures with their iphones. Including things like wood backgrounds, green grass and sky etc...a lot of the content that is ultrageneric is very, very easy to take yourself.

And yet here we are.

And if it is true that Shutterstock is getting 50% less content even if they don't take any ai but on Adobe sales are increasing although or probably because they have an additional ai collection, then ai content is our friend that brings more subscribers to the agency that pays us more and treats us well.

The easiest way to deal with ai worries is to just try it yourself. Even just for fun, you don't have to sell it.

It becomes easier to understand that it is just a tool, like Photoshop is a tool.

I'm not an experienced AI prompter, but I wonder whether that's really the case. I can imagine a lot of situations where AI prompting seems to be faster and cheaper than actually producing the image. Finding (and if needed, renting) a location, finding (and if needed hiring) the right models, propping a set, hiring a photographer (or investing your own time) and do post-production seems to be way more time-consuming and costly than paying a competent AI prompter to generate the AI image. Thinking of generic business settings here for example, or generic lifestyle situations. 

I agree that it's just an additional tool, and that there's no other way than embrace it. It won't go away, it's here to stay and to play a dominant role in certain market segments. But it also opens up the market to to a lot more people than only photo/videographers. I'm convinced that having photography skills can also improve your AI prompting quality, but it's no hard requirement. So this definitely brings in more competition (yes, just like what smartphones did to DSLR's and what DSLR's did to SLR's...) resulting in lower individual sales volumes for those who are competing in those very saturated segments.

imho it all comes down to a very simple logic. The market might still be growing, the supply, driven by technology, is even growing stronger, and some segments are becoming even more saturated than they already are. This causes lower individual sales volumes, lower value for individual content. This is what we are seeing for quite some years now, and if you ask me AI will speed up that process.

Maybe even up to a point where buyer experience is disturbed, because they will get the feeling to be looking for a needle in a haystack, but that's a different discussion :-).

Appreciate it is in grey zone but just because agencies adjust their terms of service if AI can't be copyrighted (ok debatable) why would people pay for AI images on the agencies they could just download and use for free. 

Appreciate it is in grey zone but just because agencies adjust their terms of service if AI can't be copyrighted (ok debatable) why would people pay for AI images on the agencies they could just download and use for free.

They pay for the same reason they don�t just steal content from the internet or download a file with a commercial cc license from flickr: quality control of the file by Adobe and legal reassurance.

The majority of content used on the planet is stolen, not paid.

Then there are agencies with tons of �free� content.

And then there is the cheapest option: take your own picture with your mobile phone.

And now there is ai.

The most important thing we sell is TIME, not content.

The customer can browse the edited collections of agencies and save tons of time.

High end customers can also ask for personal curations from the agency editors to save even more time. Personal service that is still cheaper than making your team search.

Content can be sourced from many places, often even for free, but a if you compare browsing Adobe to sifting through Billions? of free images on flickr�it is not comparable at all.

The marketing team employees are expensive, stock agencies are a very, very cheap resource compared to other options.

So, if Unspalsh didn�t destroy all the agencies and the billions of free flickr files don�t stop the downloads, why would ai?

Quote from: Roscoe on February 23, 2024, 08:39

You demonstrated things perfectly. The fact remains that this technological turning point that we are experiencing is truly unprecedented, difficult to copy/paste from the past. But there are also other image industrial powers outside of AI. And they may want to fight on the legal level.

Hmm not sure? Look what the internet did to everyone's lives. Look at how social media impacted our society and politics. What companies like Uber or Airbnb did to regular taxi drivers or hotel owners. And whether it's social media, or crowd-sourced service companies... all of these developments had their times of operating in legal gray zones, and most of the successful companies in these areas blatantly crossed lines. Worst case they just came away with it because they could blame someone else, or best case they had to pay a fine, which they (years later) happily did because they already made a ton of money while operating in legal gray zones.

You are right that AI is another technological development or causing a turning point, and difficult to predict how huge the impact will be, but it's not really different than other developments. It destroys previously established and seemingly robust companies, and creates new ones that might become even bigger. It takes away jobs from some people, and creates new opportunities for others. Only thing that's different is the speed, but also here: technology has always developed with increasing speed, as one tech helps out another one. We just came to a point where democratic policymaking is unable to keep up with the development, complexity and impact of new technology.

Quote from: angelacat on February 23, 2024, 08:55

Appreciate it is in grey zone but just because agencies adjust their terms of service if AI can't be copyrighted (ok debatable) why would people pay for AI images on the agencies they could just download and use for free.

They pay for the same reason they don�t just steal content from the internet or download a file with a commercial cc license from flickr: quality control of the file by Adobe and legal reassurance.

The majority of content used on the planet is stolen, not paid.

Then there are agencies with tons of �free� content.

And then there is the cheapest option: take your own picture with your mobile phone.

And now there is ai.

The most important thing we sell is TIME, not content.

The customer can browse the edited collections of agencies and save tons of time.

High end customers can also ask for personal curations from the agency editors to save even more time. Personal service that is still cheaper than making your team search.

Content can be sourced from many places, often even for free, but a if you compare browsing Adobe to sifting through Billions? of free images on flickr�it is not comparable at all.

The marketing team employees are expensive, stock agencies are a very, very cheap resource compared to other options.

So, if Unspalsh didn�t destroy all the agencies and the billions of free flickr files don�t stop the downloads, why would ai?

I agree with all of this, but some nuances.

imho free libraries like Unsplash, Flickr, or others do take away volume from paid downloads. I see a lot of content in magazines or newspapers that is licensed via one of those channels. That would have been a paid license otherwise. We don't know how big that impact is, but I think it's fair to assume that free libraries do have an impact on our download volumes.

AI is slightly different, because it's provided by the same agencies as the ones who offer non-ai content, and paid downloads, but it has an impact on the downloads for those who stick with uploading non-ai content.

And I think that's OP's point. Are we doomed? Those who adapt probably not. Many who don't adapt probably yes, or at least severely impacted, and a minority probably also no because for other reasons.

Anecdotal example. I recently came across a wedding photographer who still shoots weddings on film. Analog. He charges a few $1000's, way more than others who shoot digital, and his business is thriving. He didn't adapt and is doing well because he's very good, and he's selling his niche as an asset, a scarcity, or even a gimmick. Who knows. It works. Needlessly to say: he's the rare exception. Many others unwilling to adapt went out of business.

I know a person in India who has hired 12 year old kids from a school to generate AI images, add keywords and upload them to Adobe. He manages 16 different accounts using names of family members and is making thousands of dollars. He told me that he uploads over 1000 images a day and is trying to expand his business to upload at least 3000 images a day.

Some will say that it's the same thing as in Leonardo's time, only a brush and a canvas was needed... 

Some will say that it's the same thing as in Leonardo's time, only a brush and a canvas was needed... 

Which is kinda true? Very talented painters, with a unique style, idea and persona will still make their money with painting.
But of course, that takes way more than just a brush and a canvas.

Just like it takes more than only a digital camera and a memory card ;-)

Quote from: DiscreetDuck on February 23, 2024, 12:27

Some will say that it's the same thing as in Leonardo's time, only a brush and a canvas was needed... 

Which is kinda true? Very talented painters, with a unique style, idea and persona will still make their money with painting.
But of course, that takes way more than just a brush and a canvas.

Just like it takes more than only a digital camera and a memory card ;-)

Yes.But some here think that making stock photography is making snapshots.
For exemple, I just saw that one of my photos arrives number 5 out of more than 1.3 millions, taken a few months ago, with just one keyword for the animal it represents. Maybe the decades I spent working on my skills can still make a little difference.

Hem, maybe I should point out that the keyword I was talking about wasn't an AI prompt... 

I know a person in India who has hired 12 year old kids from a school to generate AI images, add keywords and upload them to Adobe. He manages 16 different accounts using names of family members and is making thousands of dollars. He told me that he uploads over 1000 images a day and is trying to expand his business to upload at least 3000 images a day.

Lol. Yep - many east indian "business models are "steal or spam as quickly as possible, and try and get away with it before anyone notices".

In 2 3 years from now Stock photography will be completely dead. This is so disruptive that nobody will loose time scouting a huge database when with a few promts you will get what you want and nearly free.

The only type of stock photography that will survive is editorial. This might be the reason SS is buying editorial agencies because they know the creation of images and video will be of Open AI, Google and the likes. The stock sites are irrelevant. Not Getty nor SS will survive this Tsunami. May be Adobe because they are a powerful software company that might battle the giants.
I think they have no chance against OpenAI or Google

This is game over Weird that some people do not understand that already

Perhaps, but what takes more time, what really is the time comparison between "scouting a database" (which btw will likely return strong results since it's top results are images which are desirable by virtue of sales)
and typing in a prompt and looking through all the results find a desirable image (never-mind the artifact issues)? I think ultimately you are probably correct, the question becomes how soon will it happen.

I also wonder ultimately how the cost comparison breaks out (generating AI images vs. purchasing stock)?

Even images of stuff that is seemingly available everywhere are no always that easy to copy.

For example really clear blue sky with a nice cloud distribution does not occur as often as you might think. And if you want a specific type of clouds like cirrus clouds, you may have to wait a long time until you get an opportunity to get a really nice picture.

it's not that hard - use PS sky replacement tool to add bluesky or create more dramatic images




i also take sky images on those 'rare' days and add them to my PS collection

You are assuming that every customer wants to spend time playing and learning ai.

I know a lot of people with agency plans that tried it but quickly gave up.

Especially when you really are not sure what you are really looking for and need, it is a lot faster to browse a very large database and pick out a few files.

I do believe people will use ai to adapt what they download.

You need to have a basic understanding of photography or video to even prompt what you want. If you don�t know what rembrandt lighting is, how different lenses affect an image, what color processing you might prefer etc�there are endless variations. Differences between acrylic and oil color painting or even a percentage wise mix of different art media�it is very complex.

Plus a lot of the people with an agency plan are normal people, not media professionals with years of training.

Again, those who believe it is all over, it is probably best if you readjust your income sources.

But it is not over for everyone.

Yes.But some here think that making stock photography is making snapshots.

It's not?

However, the concern for longer term remains as people start to generate their own AI stock images for their use rather than buying what we generated.  Shutterstock is going the different rout letting customers generate by themselves, but not accepting any AI images from contributors thus keeping 100% sales to themselves.  Eventually, they  won't need us uploading photos/videos.  We are being replaced.

I think you're right, we are being replaced, mostly, except for some people needing images, who won't want to learn how to make their own.

There's still room for creative work and making things, ready to download, for stock artists to create with personal skills and understanding that AI just doesn't have.

There's still space for real and editorial, that AI can't make.

But for simple stock images, illustrative, icons, decorations, accessory images, AI will cut into artists earnings. Agencies will love to cut us out as we are an expense.

Quote from: cobalt on February 25, 2024, 07:46

You are assuming that every customer wants to spend time playing and learning ai.

I know a lot of people with agency plans that tried it but quickly gave up.

Especially when you really are not sure what you are really looking for and need, it is a lot faster to browse a very large database and pick out a few files.

I do believe people will use ai to adapt what they download.

You need to have a basic understanding of photography or video to even prompt what you want. If you don�t know what rembrandt lighting is, how different lenses affect an image, what color processing you might prefer etc�there are endless variations. Differences between acrylic and oil color painting or even a percentage wise mix of different art media�it is very complex.

Plus a lot of the people with an agency plan are normal people, not media professionals with years of training.

Again, those who believe it is all over, it is probably best if you readjust your income sources.

But it is not over for everyone.

Agree, just to add one more detail on Ai generated process. It gives us many "superstar" images just in coincidence on many factors all together in the proper moment and prompt, then this great image will be available in the stock agency for cents. You may not reach such a great photo for hours of prompting while one professional who work this will have it routinely in the working process.

It's not really thought proper through to the end.

I already see several blogs and smaller online newspapers that have switched to DALL-E or Midjourney. And this is the beginning, for whom poor quality is already enough to use such images just like a gap filler.

I am convinced that several AI providers will create prompt templates in the future to make prompting much easier.
It's very unrealistic to believe that you will need to "study promt engineering".
You already can get very decent results with very easy prompts if you use the right models.
If some one would write a small introduction which model you should use, it will be a no brainer for the users.

The new Stable Cascade model from Stable AI is already taking a first step in a similar direction. Here, a text model translates the user's prompt input into more complex prompt input for the actual image model.

As soon as the performance has been improved to such an extent that images can be generated in real time at the same speed as typing, a large generic share of the image stock market will become obsolete.
That is only a matter of time. As technical developments are exponential, I would guess a max of 5 years.

Sure you can still make money with AI images. I currently "raised" my earnings to about 40 - 50 USD / month with a port of approx. 1300 images.
But it will be even with a port of 10k images just a small additional income to make small travels or to take my wife out to the restaurant.

What will survive as AI stock are really creative concepts and striking image compositions. Images, which give you a "wow that's looks really awesome, haven't seen this before".

But also photos with authentic regional places, tools, architecture, etc.
And as last also professional and very detailed collages will survive.
All stuff, that AI can't generate because there are not enough images for the model calibration.

I would say the main photo stock market is already dead.

Lots of fair point from people who believe stock is basically dead. It is important to act on it and readjust the income flows.

ai is the biggest change in creative work ever. So everybody will have their ow solution.

Quote from: Big Toe on February 25, 2024, 13:51

Editorial images are images that can only used for editorial uses, like for example to illustrate a newspaper article and must not be used for advertisment. They also should NOT be altered in a way that compromises their editorial value

I bet that's what you meant?   

Yes, thanks, that is indeed what I meant, I amended my post accordingly.