The Verge: OpenAI announced ChatGPT Images 2.5 on Tuesday and is adding a new way to tell ChatGPT what you want it to make an image of: by drawing a doodle. With a new feature called Sketch, you can just draw something right inside ChatGPT and then tell ChatGPT how you want it to make an image based on that sketch.
6 comments:
I don't think it is at all shocking that I find this idea just a little bit distressing. I understand that it could be a good feature for some people that are really bad at drawing and are trying to communicate an idea visually quickly. But I also feel that this tool defeats the purpose of drawing because it always manages to show a person's personality, and at least personally, when I’m doing simple doodles I don't really want them fully rendered out in AI. I also think that taking a doodle of a cat and spitting out a photo realistic picture of a cat really isn't all that different from generative AI images, it just involves a different sort of input. Either way it feels like a soulless recreation of actual art and like a thing that doesn't really need to exist. Instead of generating an image off a sketch it feels like it would be easier and better to just make a mood board or collage to communicate visual ideas.
I think this new chatgpt 2.5 version can be very useful and is also slightly terrifying. For artists, designers, or regular people who aren’t the best at drawing but need to convey something visually, chatgpt 2.5 is a very useful tool for those situations. It helps people convey ideas faster and clearer especially when it comes to communication across hundreds of miles. But chatgpt 2.5 is also terrifying, it is extremely hyper realistic and threatens the jobs of hundreds to thousands of people all across the world. There's a certain authenticity and "organicness" to art and images that come from real people compared to AI. There is an extreme level of controversy on this topic and I do see both sides there are benefits but I also do not think the benefits outweigh the negative impacts that come from letting AI farther into our lives.
For a scenic or costume designer, this could be a massive improvement to their workflows. Designers can start with thumbnail sketch which then the AI could rif on and upscale. Or a designer could sketch out a whole set and have that turn into a fully rendered visual. Or a costume designer could draw out a costume drawing and then have a fully detailed render they can use. Additionally, text context could be used to make it more specific.
These tools will become important in collaboration. I have talked with designers who use AI images for their moodboards because you can just request anything. It is great for getting on the same page as fellow designers.
I often hear people talking about how AI is trying to replace our creativity. I dont believe this is true. If used correctly, AI can be a tool to streamline existing workflows so that more time can be spent on the initial creative process and the final implementation while spending less time on the middle prep phases.
Not taking into account the serious problem of people losing creative skill and removing artists' jobs, I think this is a very useful and good feature. I cannot imagine how much time this feature will save, as well as improve clarity and communication.
Looking at ChatGPT Images 2.0 in a theatre setting, I can see scenic, props, and costume renderings being way more clear and effective. Additionally, designers can insert a prompt or draft a general idea and ask the feature to output am image that they may be struggling to visualize. Sitting in on multiple artistic creative meetings, I've seen designers present collages. If the designer appropriately cites ChatGBT Images 2.0, I don't see any problem with them utilizing this tool when presenting in world building, story board, or synthesis meetings. While this update is terrifying, I think approaching this tool cautiously but openly is important. The technology is only improving.
I can see why people like using AI for things like this. It makes anyone feel like they can be “creative” but I absolutely hate that. I have gone on this rant so many times, but art is accessible as is. Not everyone is immediately a good artist of realism, but art is a skill that can be trained and can be done with cheap supplies. People treat AI as a way to get “real” art without paying for it and once again AI is being used to skirt around money and take away jobs from people. I would so much rather go up to one of my friends and ask them to draw me a realistic cat and pay them for it than ask AI to generate an image. There are also plenty of ways to get the image that you are asking for. There are so many sites that have collections of photos that you can use and you can photoshop some together to get the exact image that you want. It just takes a skill that a lot of people believe is not impressive/necessary,
This article get two sets of opinions and emotions that out of me, and they are both centered around weather or not the images being generated are being used for financial gain or not. At face value this ability for an AI to generate a detailed image from a sketch can speed up processes in making artistic displays, but at the same time by using an AI to generate art like this you are taking jobs away from the artists that would have been the only option to create the art that people are now asking AI to do. So to summarize I like how easy AI can make any art that you want based on a basic sketch, but I am sorry that this process is removing human creativity from the whole situation. In a way I also look forward to and fear the day that it becomes impossible to identify the differences between a piece of digital art that was created by an AI or by a person. As once this happens people everywhere could get away with advertising AI generated art as an original piece of art made by the person who presented it.
Post a Comment