Help Us Shape the Future of Jewelry Design! We’re a team from the University of Pennsylvania building an AI-powered 3D jewelry design tool that lets you generate photorealistic models from just text or images — perfect for custom design and e-commerce.
If you’re a jewelry designer or manufacturer, we’d love your feedback. This quick 3-5 mins survey will help us understand your needs. You can also sign up for a paid 30-minute interview ($25 gift card) at the end.
I get that what you are working on is exciting to you, but have any of you considered inevitable unintended consequences? They can be significant.
Example. 1948-50 - few if any had air conditioning, few had a TV, and no one in my neighborhood did. When summer came around people would sit outside in the evening with their neighbors, maybe play cards or just chat while we kids ran around the front yard catching fireflies, etc., enjoying being able to stay up late, knowing we were among friends and safe. It was very nice and you can’t buy that experience. People knew their close neighbors and were friends with them. And I’m talking New York City, not Peoria or some very small town. It was very nice and you can’t buy that experience.
TV killed that. Everyone stayed home. People stopped going to dance halls and bands dried up.
Most people are fine with all that but there was a breakdown of community. I like my neighbors, but they have their life, I have mine. We say hello, wish each other well. that’s it. I don’t know them, they don’t know me. But they know what will be on TV tonight.
Now you want to have fake images. Yes, I know you think you are empowering jewelers. But generated - i.e. FAKE images. Yesterday I saw a convincing image of Donald Trump as pope. It was photographically realistic, and at the same time it was totally rubbish and extremely offensive. A ‘blessing’ of technology.
It used to be said, ‘seeing is believing’. Well, thanks to AI folks, I no longer trust a single damn thing I see online. For example, JCK Online shows a lot of jewelry pieces, many of which have ‘diamonds’ without any hint of sparkle, just gray and white, and where the polished metal does not show any reflection of a camera. I have to assume they are renderings - i.e. FAKE. Fake does not impress me, I do not believe JCK Online images.
I’m sure you are going to go ahead with your project because you code and you enjoy the challenge of your project. But make no mistake about it - MAKE NO MISTAKE ABOUT IT - you are going to KILL trust in what people see. I’m sure future generations will adjust perfectly, never having developed any trust in what they see to start with, but is this really an outcome you’d feel good about? Others may have a different opinion, but I think the cost of what you are doing will cause a lot more harm than good. Already I do not trust images I see online, and as they get better I will trust even less.
Are you o.k. with that? Thanks so much.
Neil A
I have been making twisted, forged and fabricated silver and gold jewelry for over 50 years. My designs come, in part, from those created by our father who taught both me and my brother how to make Dad’s somewhat unique style of jewelry. Our designs are based, in part, on the South Western style that influenced our father when he first learned in the mid 1940s. His style continued to evolve as a result of moving to the North East and just very practical influences like the forms of raw material that we could buy or make, the cost of the raw material and the desires of our ever changing market. Over time, both Don and I have followed our own creative paths. Our work, while it reflects the influence of our father, has evolved in different directions to reflect our own interests and creativity. In my professional life I managed all aspects of instructional and administrative technology for large educational institutions in NY. I am well versed in the nature of data, how to create it, manipulate it and use it to support or refute whatever you might want to influence. To me, the internet is just a very large and sophisticated database. I have never worked with AI, but as I hear more and more about it, it seems to be as much a database as anything else. I would suggest that, if someone wanted to design jewelry using AI, they ought to first have a hands on experience with the medium. They should learned how metal behaves when they try to move it, join it in some fashion, cast it, use it to display organic and inorganic objects and most, importantly be worn in a safe fashion. From a practical point of view, it should be designed in such a way that someone could make it, manufacture it, fix it and wear it. I am sure that AI will, good or bad, makes its way into the process of designing jewelry. I would just suggest that the people populating the fields of this very large database have some first hand experience with the process of working with metal and somehow digitize these experiences. If the data that drives this AI just comes from a review of the final product as it is displayed in the media, they will miss that which makes a piece work or not work. Good luck with your adventure and keep is posted. Feel free to ask questions as I am sure that you will have Many…Rob
My suggestion is to take a real live butt in a jewelers chair, learning the trade and it’s little nuances. After decades of making, I’m still learning different tiny little nuances that can make or break a jewelry piece. This from the prospective of a Jeweler and being married to a former VP from IBM, to now he is one of the top IT pros at Disney. A few survey questions are not going to make a good program to encompass the field. Learning before trying to generate an AI program to do what decades of experience can do. Sorry if this sounds mean and hurtful, but reality is still experience before you can program another program to do what we do.
When CAD/CAM first started I noticed that quite a few of the designs didn’t make sense from a metal fabrication point of view. Strange floating parts were plugged in, ring galleries didn’t flow, curves were awkward, etc. I noticed that the CAD designs that looked good were done by skilled jewelers who learned CAD. They understood how to make the pieces and used CAD as a tool. For instance, they built their ballerina rings in parts. They cast and polished and pierced the skirts and other the parts (shank, gallery) before soldering them together, just as they would had they fabricated the whole thing. Enough heft was built into the design to withstand the polishing and fine tuning process. Prongs for center stones were soldered in wires instead of cast ones.
I am concerned that bench skills will die out because people want to jump into CAD and skip the arduous, time consuming skills development. It shows when the aforementioned ballerina ring is designed in CAD as a single piece and cast. They aren’t finished inside and out. I have seen them with flimsy, straight gallery wires instead of gracefully curving ones.
I am sure that many of you have noticed the scarcity of skilled watchmakers. Fine watches have to be sent to Europe to be repaired because the number of people who can actually make replacement parts are unicorns. It would be sad for this to become the case for bench jewelers.
I am not condemning all technology. Lasers are a godsend. They allow repairs to be done and
are definitely a fabulous tool in jewelry making. Just being able to tack things together to solder is a time saver and increases accuracy. Connecting different kinds of metals with different expansion characteristics or just connecting pieces without annealing them is so handy.
I guess what I’m getting at is that new technology is useful, but I don’t want it to come at the expense of valuable and treasured skills. I’ll get off my soapbox now.
Donna
I believe that AI is here to stay and at the same time it is here for me to largely ignore at least as far as fabrication and design is concerned in an artistic pursuit. I have said before that I don’t look at other peoples work. I do look at other work, however. (I study Celtic jewelry, I don’t look at Seamus McCool’s jewelry. I look at Native jewelry but I don’t copy Ben Nighthorses work) I want the work I produce to be my work generated by my idea of what a pair of earrings should be, based on experience and study but with little outside influence. A fine line to draw I know. As Rob has said we both do our Dad’s designs and we both have our own designs. We have our own techniques and methods of production and we are both aware of how metal moves and works in our hands. It is those understandings and techniques that makes our designs unique. AI can look at a penannular brooch and create a program for a 3-D printer and produce a model that could be cast in silver, gold, or pewter. The end result will be one or one hundred brooches that will be absolutely the same. Where is the art in that?
If you are trying to create an industrial process to make mass amounts of the same piece of jewelry I suppose AI is the way to go. But if the thought of an artful effort has even been a part of your consideration then start at the beginning with a bench and some tools and someone willing to help you learn the metal moves. Learn how what degree of heat is not enough, just right, and too much. But first put in the time it takes to learn how to create the art of jewelry making.
Don
You seem rather unaware that artists/artisans, such as those who create this wonderful forum, are the people making the work that AI is stealing and regurgitating. I know this is just the way it is now, but I don’t think you’re going to have all that many creative people excited about this. The people who will purchase this service are those who have no interest in learning the skills, and little interest in creativity. They will just like the idea of selling jewelry for whatever reason… Maybe it sounds glamourous if you don’t have to have filthy fingernails and a sore neck. 
It seems like what you’re creating here is an infringement machine. The sorts of decisions that lead to a unique and beautiful jewelry design are not susceptible to expression in a word prompt. You might say “ornate” or “symmetrical” or “colorful” but that won’t tell the ai engine to do anything special. What people are going to ask it will be more like “Give me a platinum and gold bracelet in the style of David Yurman, but not exactly like any of his actual designs,” If you’re enabling that, don’t be surprised if you find yourself in legal difficulties. Aiding and abetting copyright infringement is a crime. I doubt that pleading that you’re just an academic researcher will be enough to get you off the hook. Court Confirms the Obvious: Aiding and Abetting Criminal Copyright Infringement Is a Crime – Center for Intellectual Property x Innovation Policy
I took your survey.
I have questions and I have suggestions.
CAD and AI jewelry designs only work in “real life” when the designer has had hands on training. It takes a good 5 years just to learn the basics and another 5 to master the art of designing and making jewelry that will stand the test of time.
So my question is … How many of you at U of P have actually hand made jewelry? When making jewelry, a strong background in practical engineering, is absolutely critical. As well as geometry, math, metallurgy, and chemistry. Can you imagine a city asking an AI designer with no engineering skills to design and build a bridge? How well versed are those of you in the AI biz in those skills?
My husband and I do bespoke high end jewelry for the “Invisible Wealthy”. They do not buy "Brands. When we sit down with clients we ask questions of what they want. And then we do hand renderings with pencils and paper that look realistic and 3 dimensional right in front of them in 15- 20 minutes. When working with CAD and AI clients don’t receive a rendering until days later. Not to mention the extra expense. Statistically our “close the deal” rate is 99 times out of 100. Our pieces last a lifetime and longer. The current trend of “minimalist delicate jewelry” is pretty much all CAD driven and the mountings are failing, breaking, and loosing gemstones at an alarming rate. Not a good biz model for repeat customers and their trust.
Pretty much the vast majority of professional trained jewelry makers and designers will likely have the same sentiments about having a good foundation of skills.
[