Ars Technica: On Wednesday, Midjourney announced version 5 of its commercial AI image-synthesis service, which can produce photorealistic images at a quality level that some AI art fans are calling creepy and "too perfect." Midjourney v5 is available now as an alpha test for customers who subscribe to the Midjourney service, which is available through Discord.
5 comments:
I've seen a lot of AI generated content the past few months, most of which is harmless. I've also seen a lot of buzz surrounding the harmful uses (ie, all the people using chatGPT to write their essays for them, even though most AI generated writing has an immediately recognizable tone). Seeing these images makes me wonder what harm this may cause to those seeking roles in graphic design and photography. Surely AI generated photographs will rarely win contests or be displayed in a museum, but the type of visuals people commission for websites and advertisements seem at serious risk of being replaced with these AI generated tools. On the other hand, Pandora's box can't be closed. There's no preventative measure anyone could take at this point to change the possibility of people leveraging these tools to displace creatives. However, public access to AI tools like Midjourney also puts creativity in the hands of new people, the way photoshop did, and the way cameras did. Just like how painters transitioned from aspiring to heights of photorealism pre-photography before leaning into a new wave of abstract expressionism, graphic designers and photographers will surely have the opportunity to explore new territory as technology once again fundamentally alters the artistic landscape.
AI art has been such a big and trending topic right now. For the previous months, one of the biggest debates when talking about AI generated human portrait is the hands. The progress that it has been making in creating a perfect piece of art is so scary but at the same time so fascinating. The amount of changes and improvement that has been done in such a short amount of time is keep on reminding me what technology is capable of doing these days. As it keep on improves, it keeps inspires and makes me think about if using these technologies in my design is capable. If there is such tool where I can put the human measurements in and it would give me a rough model in how the personal would look like, it would help my design process in costume and especially if I can put my design on the human model to reference off of.
I like AI generators as a tool for fun or for mood boarding but the moment AI generated images or videos are used to publish works that are harmful to the creators or people in the photos or claimed to be their own original work, I do not like it. AI generators like the ones mentioned in the article use huge databases that could only come from sourcing the internet taking artists and photographers works and using them as data to generate their images and styles making the produced images not completely an original piece and with no way to credit the original creatives involved. The fact it is becoming harder to identify if something is AI is scary, but I also think that there should be disclaimers on AI produced images to let people know the sources and the fact its machine generated.
I have heard of the discussion of AI generated artwork and music, but I did not realize that it also was happening in the photography industry. With capitalism’s emphasis on decreasing costs in order to increase profits, the use of synthetic voices seems inevitable, just as EZ-Pass has replaced human staff at highway tolls. But, currently, AI produced work is just not the same as that photographed by human photographers. These programs are more sophisticated- I would not have been able to guess that these photos were generated by AI. Since it is apparent that AI generated work will be a part of the future, clear indication that the work is AI generated must be included. This is not okay. Authorization needs to be upfront and in fact, I believe consumers should be informed when viewing photographs by synthetic artists. Companies should not be able to exploit artists’ work without compensation.
I didn't know about the details of why AI struggles with hands, but it’s fascinating. Midjourney is by far the best Artificial Intelligence prompt image generator I’ve seen so far, and it’s true that the hands look really, really good. It’s so exciting to follow the development of artistic AI, especially because of how fast it seems to be moving. It feels like just at the beginning of the semester AI was struggling to generate even the simplest things. It’s just jaw-dropping how good Midjourney v5 is compared to the past models. I am still very interested in hearing about the legal problems that have come up earlier this year, because I haven’t heard anything else about them. If art generators like these are still causing intellectual property and copyright problems, I hope they’re working to find a solution soon. It seems like such a big problem, yet no one addresses it.
Post a Comment