The Results of AI Images

Back in July 2023, I took a few days away from work and drove down to Monterey, California. While there, I was able to relax at the Del Monte Beach, ride the trolley past Fisherman’s Wharf and Cannery Row, and spend a few hours at the Monterey Bay Aquarium. I took my time viewing the creatures in the various exhibits and especially enjoyed the various displays showing creatures only found in the dark recesses of the ocean ( with the exhibits dimly lit as well). My favorite photos I took had to be of the jellyfish exhibit and some of the fish found in the Pixar movie, Nemo (shown below).

Colorful fish swimming around coral and anemone at the Monterey Bay Aquarium. — Photo by Beriah Grier

While doing a search through Creative Commons licensed images, I was able to find one that resembles the one I took. The aquarium exhibit pictured is a lot larger and more populated than the ones I saw at the Monterey Bay Aquarium.

“Coral reef and colorful fish from an aquarium” by Marcus Burnette is licensed from CC0

As in my weekly discussion assignment, I chose to use Google Gemini as my interface of choice to generate a similar image to mine and the Creative Commons licensed one. I told it to generate an image of colorful fish swimming around coral and anemone in an aquarium. From various searches I did, it would come up with a more realistic camera photo of a close-up of one fish, and the rest of the photos looked more like cartoonish, painted photos of schools of fish. I tried the prompt a few times, emphasizing a realistic camera image and a realistic photo, but the images still came out very stylized. I changed the prompt to generate a camera photo of colorful fish etc. and those images turned out to be a little bit more of what I was envisioning (not 100% accurate as displayed below).

Gemini image of fish in an aquarium — Google Gemini rendering of fish in an aquarium, created using Open AI

While I can look at the photo directly above and know that it’s fake because of my research, if this was just placed on the internet without any context, I might not tell the difference. This raises the following question: Can we trust the images and videos that we see on the internet? As this CNET article notes, there are major positive effects of using AI to enhance photos, but deep fakes and other issues will continue to become more prevalent as AI is enhanced. That is why it is important for media users to practice discernment and media literacy as they navigate the web.