We try to answer a few questions stock producers might have:

• The current state of stock: Is revenue going down?
• How many photographers are switching to AI content?
• Does AI content sell?
• Can I expect significant income with AI content?
• What are the successful keywords for AI content?
• Will it kill the stock industry?

Link to blog post: https://www.stockperformer.com/blog/is-ai-killing-the-stock-industry-a-data-perspective/

We try to answer a few questions stock producers might have:

• The current state of stock: Is revenue going down?
• How many photographers are switching to AI content?
• Does AI content sell?
• Can I expect significant income with AI content?
• What are the successful keywords for AI content?
• Will it kill the stock industry?

impossible to answer most of those questions for stock in general, much less a newcomer like genAI.

This is slightly off topic, but I wanted to comment on your introduction as I think it's missing one very important element:

"When Stock Performer started in 2010, there were a lot of complaints that the stock industry was dying. The rush of the agencies to implement subscription models, triggered mostly by Shutterstock�s success, led to a lot of uncertainty about photographers� income. After all, customers were able to buy images for mere cents now. Talks of impending doom was nothing new, however. When microstock, cheap equipment, and the internet entered the scene a few years earlier, everyone earning a few hundred dollars per sale before certainly thought it was going to be the end of their career."

You then continue to talk about the introduction of AI as another round of similar worries about whether there was a future in supplying content to stock agencies.

I think there was a fundamental difference between the introduction of microstock - primarily iStock and Shutterstock, but to a smaller degree Fotolia (they pioneered offering sites in languages other than English and sales in local currencies) - and any of the subsequent agency shenanigans in the market for stock licensing.

Microstock - web sites licensing royalty free with immediate downloads at lower prices - brought many new buyers into the marketplace for image licensing. It wasn't just lowering the price for existing buyers, but all sorts of businesses who would never have been able to afford to use images could now include photos and illustrations. Don't forget that initially, just about everyone involved in the traditional stock business viewed microstock as cheap rubbish and not really competition for them, but there was a lot of bog standard content (apple slices, shots of well-known cities...) which anyone could now deliver. Microstock did change things for traditional stock agencies - rights managed licenses are complex for customers to deal with and small businesses didn't have $500 per image to spare when setting up a web site. Now buyers had options.

Gen AI content isn't, as far as I'm aware, bringing new buyers into the market, just offering new content to the people who've been licensing stock for the last couple of decades. Stock licensing can be very trend-driven, so one possibility is this trend has its moment and then we get something else in a year or 18 months. I read recently that film photography is now a trend for weddings, so possibly traditional, human-created photos and vectors will be a new trend to follow on from AI?

Shutterstock's pairing of "margin optimization" with building the biggest collection (largely abandoning quality metrics IMO) seems short sighted and mostly shareholder (not customer) focused. The problem of reducing per-item price and betting that you'll make it up in increased volume, is that at some point there are no more new buyers and no more increases in the need for images and you can't make the volume go up enough. They've also squeezed contributors pretty hard so there's not much left there.

When agencies are focused almost exclusively on swiping customers from the competition, that doesn't help contributors. It might if the agency winning at the customer grab paid contributors more than those losing customers, but that's not the typical case in the last 5 years.

I've been pleased to see a big increase in my sales at Adobe Stock and I'm guessing (I have no data to back this up) that they're doing a good job of enticing Canva customers to their various Creative Pro offerings. But looking at the big picture, if we don't have new customers and we have gobs of new GenAI content available for licensing, that will put downward pressure on contributor royalties.

I worry about agencies paying lip service to the rights of copyright-holding creators in their dealings with AI (but that's been bothering musicians, artists, writers and others too). I see problems for contributors as largely the business models of the agencies, not AI per se. Royalties to creators are an expense shareholder-focused businesses want to minimize. Anything non-royalty-bearing that customers will buy is a win in that scenario.

Then there is the time factor,many do not have time to insert and trying different text to look for an image suitable for their purpose,many customers do not really care in my opinion,because stock contents are cheap anyway,you see the content,you like,you buy,don�t need to generate anything.

It doesn't take longer to enter what you want into an AI prompt bar than to enter it into a microstock search bar and you get a suitable result instantly. Only if you want something very very specific you might need to "insert and trying different text", but the same goes for microstock and real photos. If you want something very specific, you might need to go through 100 of pages of search results (and possibly not even find what you are looking for).
I don't get why people keep thinking it's easier to use a microstock search funtion than to describe what you want to an AI instead.

I'm seing Shutterstock accounts, which pulled off >200.000 a.i. generated images within a few months. How can RPI be better compared to a conventional artist adding like only 500 images per year.

Can you show us the links to these portfolios? Would be interesting to see what kind content they generated.

Quote from: Thomas Vogel on May 29, 2023, 09:26

I'm seing Shutterstock accounts, which pulled off >200.000 a.i. generated images within a few months. How can RPI be better compared to a conventional artist adding like only 500 images per year.

Can you show us the links to these portfolios? Would be interesting to see what kind content they generated.

here's one example: shutterstock.com/g/agsandrew

i could crap out hundreds of such images, daily ...but where's the sense when there already a gazillion similar a.i. images there?

Quote from: leremy on May 29, 2023, 20:01

Quote from: Thomas Vogel on May 29, 2023, 09:26

I'm seing Shutterstock accounts, which pulled off >200.000 a.i. generated images within a few months. How can RPI be better compared to a conventional artist adding like only 500 images per year.

Can you show us the links to these portfolios? Would be interesting to see what kind content they generated.

here's one example: shutterstock.com/g/agsandrew

And where does it written that this content was created in AI?

It actually looks too simple to come from ai. Midjourney especially would generate much better images.

Not saying the person is not using software, but it is not a high quality look and with ai that would be easy to do.

I am not an illustrator, so�but just by looking at the Adobe collection, this does not feel like gen ai content.

Quote from: stoker2014 on May 30, 2023, 09:50

And where does it written that this content was created in AI?

Hmm, maybe this person is just a genius, being capable of imitating the style of hundreds of artists and spewing out hundreds of variations of every of his ideas. But he doesn't seem to be capable of creating high resolution content.

And why is it not written anywhere on shutterstock that this content was created by AI. After all, shutterstock should write it.

Quote from: Thomas Vogel on May 30, 2023, 07:15

... I particularly refer to images of artists that have been used as training material without their consent or licensing, essentially representing the brain of the AI. Technically, parts of these artists' images are processed with each new prompt during image generation. Personally, I would only make my images available under an extended license for generative AIs....

like many othersd, you fail to learn how genAI works before criticizing..

 'parts of images' are NOT used in creation of new art.

Like many others, you failed to understand copyright, which says, "Don't touch my images unless you pay for it." Downloading and processing by any software is clearly prohibited by copyright.

'parts of images' are NOT used in creation of new art.

Sorry, I didn't read it right. I wasn't speaking of "used" as in "take that ball from my picture to use it elsewhere". I was speaking of "processed." To clarify this: Even if a prompt and the resulting image have nothing in common with my image, my image was still part of the processing with every incoming prompt. In the same way, all the other million images are part of the processing, all at the same time. Processing, such as "Oh�Thomas's image is not the direction we want," is still a processing, since it helps to guide the AI in the right direction.

Hi Thomas,

I understand it looks crappy and mass produced, but I thought this type of content doesn't need ai?

Perhaps I misunderstand illustrations, there have always been ports with thousands of very, very similar images, long before ai. How did they do it?

Whatever software he is using, it is not very useful for customers.

But because the 2000 similar files are all competing with each other, I doubt he is having a lot of sales and the individual files will have a very, very difficult time rising in rankings.

Stock karma is a bitch.