Research Study on Generative Fill for Class

I am conducting a research study (for a graduate course) on how age, experience with Photoshop, and level of expertise in Photoshop impact a user’s experience with Photoshop Generative Fill.

I want to understand how these different factors explain the variation in user opinion on what and to what degree the technology is or is not useful, easy to use, and aligns or does not align with user values.

I am looking for Photoshop users who have used Generative Fill and are at least 18 years of age.

If agreed, we will schedule a 30-45 minute interview with you via video chat. Your interview will be audio recorded. After the interview, you will be given a questionnaire about opinions on specific topics.

Contact me (Ian P. Swift) at iswift2@uic.edu if you have any questions.

Not interested in a video chat.
Just going to point out that generative content is now an included feature in Photoshop.
It is a tool.
One must learn to use the tools of the trade. Neither age nor experience should impact the user’s experience. Only available bandwidth (some content reference downloads will take forEVer on a slow internet connection.)

It isn’t so much about user values either. More the clients’.
If a client tells you to avoid using AI content, you don’t use those particular tools, which with Photoshop would be really hard because you’d need to determine if, for instance, using the new Quick Actions all constitute AI - like Remove Background or Refine Hair selections et al.
If the client doesn’t care, are you going to save billable time? Or not? Up to you.

A real life example
I recently bought a streaming concert ticket on Kickstarter. The concert comes with a T-shirt add on. Kickstarter now requires artists to declare whether or not AI is used in creating the items they offer for sale. This artist’s AI declaration said: (paraphrased) " The concert is 100% human. The t-shirt, because the tools of generative content are now included in quite a number of illustration software, does have some AI generative content. "
It’s now up to the buyer whether they want to purchase that add on.

1 Like

I’m uninterested in a video interview, but I have a few thoughts.

Photoshop’s generative fill feature is a narrow concern in light of other readily available AI image manipulation tools.

I worked at a newspaper in the 1980s and '90s. During the '80s, we routinely airbrushed almost every photo to make the terrible-quality Associated Press wire photos more legible and to help compensate for the poor quality of high-speed, high-volume letterpress printing on newsprint. However, the retouching often went further than that when editors requested removing objectionable items from the photos, such as utility poles behind people’s heads, jet contrails, cigarettes, etc.

There was a dramatic shift in retouching philosophy in the early '90s, a time when digitally scanned photos and Photoshop began to replace the old wire-transmitted photos. This coincided with a widespread transition from letterpress printing to higher quality offset printing and cheaper 4-color process reproduction. These changes prompted photojournalists and newspapers to reevaluate the practice of retouching photos. Where once all photos were retouched, it soon became taboo beyond basic color/brightness correction and minimal dodging and burning. This was a monumental 180-degree turn for the industry.

The reasoning was that journalism depended on people believing that the reporting was factual and that any manipulation of the photos, no matter how trivial, could undermine the believability of the reporting. Of course, the supermarket tabloids didn’t buy into this new way of thinking, but that’s another story.

Today, the issue of photo manipulation is more complex. Much coverage, for various reasons, has become highly polarized, politicized, and biased. However, photo manipulation remains forbidden in mainstream news coverage, particularly now that AI has made it so easy and social media fakery is rampant. The recent incident involving Kate Middleton adjusting her family’s Easter photo serves as a prime example. As soon as these minor alterations were detected, most news services swiftly removed the photo from their distribution channels. The use of Photoshop’s generative fill at most mainstream news outlets would be grounds for immediate dismissal.

As for me, I’ve been out of the journalism field for 24 years. Instead, I’ve held various positions in design, advertising, magazines, and public relations. In these areas, the use of photo manipulation is more nuanced. For instance, if I were designing a promotional brochure for a real-estate housing development, I wouldn’t alter any photos of the development in ways that deviate from the truth. However, if I needed a generic stock photo of a house for a brochure on rising costs, I wouldn’t have reservations about using generative fill (or other AI) to alter the stock photo to better illustrate the brochure’s content. It’s a matter of weighing whether the photo is intended to document reality or is merely an illustrative accompaniment that won’t be misconstrued.

Hi all! Thanks for this initial feedback. I’ve received approval from the university IRB to make this an official research study. Additionally, I’ve heard what you’re saying about not wanting to do video interviews. I’ve created a link for a single anonymous questionnaire that you can fill out if you meet the criteria listed in the questionnaire. Also, please note the consent form at the top, as since this is Human Subject Research, participant consent must be confirmed.

I’ve included the link in my profile.

Thanks again so much!