Is this End of Shutterstock, Adobe Stock, Freepik?

I wanted to discuss what is the future of these platforms.

Will companies, and agencies buy stock images or will they produce their own since it is much easier now to create these images?
Shutter stock removes images made with AI, adobe stock supports AI ?
I am a Shutterstock and freepik contributor and really worried about what is going to happen.
I am really confused, what do you think will be the future of stock images and videos?
Should we continue to create and submit to these platforms with original creations or adapt to the new technology and maybe start creating our own stuff instead?

I wouldn’t say so.

One of the advantages I find for stock sites is finding similar style images or charts/graphs/infographics for use of quicking interchanging the data.

For example, I might find a particular style of graph to be interesting and I then search for your name and other graphs you might produce and use those and interchange the data.

I don’t have to spend hours constructing graphics when it’s already done for me.

Finding similar style graphics for a publication is a challenge and there’s a huge advantage to finding a style that suits and everyone agrees on which makes it easy to use for publications.

With AI you might be able to get a 90% type of image created that you think looks good - but it requires that extra.

Here we fall into untrained eyes scenario - where the person creating an AI image might go - hey this looks really good.

But working with an artist or with a designer gives the opportunity for feedback and improvement and more detailed experience.

I think AI is here and we need to embrace it. Once more as an imaginative tool like any software.


Have you considered using it yourself to generate assets for stock sites?

Could it speed up your workflow?

You could alter the image and make it ready - and how would they know if it was AI-generated or not?

Could you take a style of some icons and ask AI to generate more versions in a similar style - fine tune them and upload?

Until the copyright issues of AI in photography and art are sorted, we can’t use it for a lot of things.
Until trademark and copyright issues are sorted in the ‘design a thing’ AI field, I’d be leary of that product too.

Photos of people, places and things are still going to be needed.
It’s going to hurt an already hurting illustration industry, but custom illustration is still needed.

Like Smurf said, time is money. Do you spend time trying to come up with the right set of prompts to get the art you need, or do you go to a stock company and get it fairly quickly?

Much easier? It is nowhere near easy enough to AI-generate marketable images for at least 75% of stock-photo applications that I can think of. As for more abstract needs, it’s too subjective to judge. A trained eye will spot AI-origin easily, and for a time it will be novel and admired. But not long from now, after 8-year-olds figure out effective prompting, the graphic designer-turned-AI prompter will draw eyerolls as the new “plugin jockey” making gamer sig art.

1 Like

HB, I think you discount the ability of the general public to spot AI-origin art
The only eye-rollers will be good designers that know the difference.

Didn’t think I was doing that PD. Was thinking more about the OP’s concern that the stock contributor business will dry up. I doubt that.

Take Flyfly, Adobe’s new generative image creation tool that’s in beta testing. Adobe says it will only learn from open-source, copyright-free, or Adobe-licensed images. I read they plan to integrate it into some of their already-existing software, such as Photoshop and Illustrator.

Adobe plans to compensate photographers and illustrators as images get used from Adobe’s stock platform as part of the AI generative process.

In addition, from what I understand, designers will be able to load into Photoshop their own photos, illustrations, and stock photos they’ve obtained from various sources. For example, I might buy an image of a mountain lake scene on Shutterstock. It’s a great photo, but I need it to look like fall instead of summer, and I need someone fishing in the lake, so I use Firefly to do this.

I think you might also be overestimating the ease of creating an AI image instead of finding one on a stock service. I think you might also be overestimating the quality of AI images generated from scratch.

This will likely change down the road, but currently, tools such as Midjourney aren’t really up to the task. Suppose I need a generic fantasy/sci-fi illustration incorporating a face. With enough messing around, Midjourney can generate one that will probably work. However, if I need actual imagery of people backpacking in the Grand Canyon, no AI tool will give me that authentic photo or video, no matter how good it gets.

A greater immediate threat, in my opinion, to the value of uploading imagery to stock sites is the competition. When the stock sites contained a few thousand images, making a bit of money from adding photos was easy. Now that these sites have millions of images and continue to grow by leaps and bounds, people are increasingly finding the images they upload lost in the crowd.

@Just-B I agree that authentic photography for specific use cases (you mentioned people backpacking in the Grand Canyon) is difficult for AI to produce, but I have to say, I have seen some examples and articles on MidJourney’s latest version 5 iteration and some of the images they showcase show huge improvements, and it does make me wonder if within the next 2 or 3 years, even your Grand Canyon backpacker image may be easy to generate something that feels authentic.

I failed to make clear what I was getting at with the Grand Canyon example. Often an actual photo is needed for authenticity, not a generated photo. Feeling authentic wouldn’t be good enough in these kinds of instances because it wouldn’t be an honest depiction of the actual situation described by the accompanying text.

For example, if I worked at a magazine and needed a photo of tornado damage to accompany an article on tornadoes, a realistic AI-generated image wouldn’t work. I would need an actual photo of tornado damage, not AI’s interpretation of tornado damage. Doing otherwise would undercut the magazine’s credibility and honesty. No matter how realistic-looking the AI-generated image might be, that image wouldn’t be useable unless it was clearly labeled as a simulation.

Journalistic integrity has been going down the tubes for years already. That’s a small step.
I recently saw a photo used to represent fossil fuel power plant pollution. Buildings with stacks smothered in ‘smoke.’
It was actually a power plant in regulation-stringent Germany on a cold January day. Condensation vapor, not smoke.

I don’t think you’re asking the right question. I would ask, “What will happen to my sales when each of the stock sites becomes flooded with half a billion new images?”

I agree that news journalism has changed in many ways. It’s become much more competitive, and individual news media outlets report what their specific audience segments want to read. However, not every news outlet is FOX or its equivalents on the other end of the spectrum or supermarket tabloids. In addition, lower profit margins have resulted in staff cutbacks and a decline in fact-checking, like mistaking steam from a smokestack as smoke. The legitimate news outlets I’m familiar with would fire a reporter who deliberately faked a photo and passed it off as the real thing.

Remember, I spent 15 years in journalism, my wife spent 20 years at another metro newspaper, and most of my friends are still writers, reporters, and editors. Today, for example, I have a conference call scheduled with a CNN editor about a writers’ workshop in which he’ll be participating and that I’ve been hired to help promote. I’m more than casually familiar with what’s going on in the news business, and although it’s changed, the public has an incomplete picture of those changes and what’s driving them.

Most news reporters I know are more concerned with how AI might affect writing and editing. For example, a reporter could collect the basic who, why, where, when, and what information for a news story as a series of notes, as they’ve always done. But instead of writing the story from those notes, AI could take those notes and write the story for them while fact-checking available details and checking for style and grammar errors.

But publications include more than the news business. They also include scientific journals, business magazines, medical newsletters, travel publications, annual reports, various trade journals, documentaries, academic publications, textbooks, etc. Whether or not and how to use AI-generated imagery depends on the publication. However, there’s still (and always will be) a huge role for genuine photography and video rather than faked imagery.

Some thoughts:

AI isn’t necessarily free.

Can you hold someone liable in real life if you get sued for copyright infringement because AI made it too similar to an artists work?

Can you buy exclusivity with AI?

I am sure soon there will be brands or labels for „organic" AI free imaging because if almost nothing is real anymore people will long for authenticity. They already begin to in other areas.

I was recently thinking about that very thing in terms of illustration.

Readily available and cheap stock photography in addition to digital cameras killed much of the custom illustration business. Instead of paying an illustrator for a day’s work, there was probably an inexpensive or free photo immediately available that would work.

Image-generation AI tools might get art directors and designers to consider using AI illustrations where they previously might have used stock photos. As people get more used to seeing illustrations, I can imagine exactly what you described: actual artist-produced illustrations going up in value as being seen as the real thing instead of an imitation.

Similarly, natural diamonds and emeralds are still worth more than the synthetically produced versions, which might be more perfect, but have no character or history behind them.

Stock images, particularly from iStockPhoto, were a Godsend for my clients. Many of the stock photos I used in my work would have cost my clients an arm and a leg if I had to hire a professional photographer, search for appropriate models, find desired locations, as well as secure permissions and releases that would burn my client’s budget to a crisp in the process compared to stock.

Call me old-fashioned and out of touch, but as far as I am concerned—Long live stock imaging!

On some Discord groups and Reddit (and possibly even DeviantArt but I haven’t checked in there lately,) there is already a trend for the “I am a Human Artist” designations. Small guilds or societies are springing up in that area.

I wasn’t even considering Fox.
More like Dan Rather and Brian Williams…and a host of others from newpapers, like the New York Times, fired for lying, fabrication and plagiarism. Key word is fired, though not until after the damage was done.
That photo of the German power plant was found in a prominent Boston Newspaper.

I’ve tried, many times, to generate stock imagery for my design projects using Midjourney, and have failed miserably.

I find it a lot easier just grabbing something from Freepik, Unsplash etc instead.

Having said that, who knows what things will be like in a year or two.

I definitely could see where you are coming from given how these tools will only going to get better as times goes on. The potential of being able to economically generate one’s own life-like images in high-res for mock-ups, posters and so on will greatly disrupt the stock image market.

That being said, looking for current stock images at the moment are still more convenient. Like Smurf2 said above, one might spending possibly hours to generate an image that technically fits the desired composition, color palette, setting, mood, yet still requires additional manual touch-ups.

In conclusion, stock images companies will continue to chug along but only time will tell how things will pan out.