Artificial intelligence (AI) is not a futuristic concept in the gem and jewelry industry; it is already reshaping the way we cut diamonds, analyze gemstones and create designs. Companies at the forefront of this transformation are utilizing AI to enhance their precision, efficiency and creativity, improve their stones’ yield, and develop more cost-effective solutions.
Navigating the 4Cs – and beyond
Diamond manufacturing firms often operate on wafer-thin margins because they purchase rough stones up front from mining companies but offer the polished diamonds to their customers on credit. That credit period can range anywhere from three to six months, putting enormous financial pressure on their operations.
Similarly, pricing for polished diamonds can differ even when two stones have the same cut, clarity, color and carat weight, since their optical performance (brilliance), spread (perceived size), optical symmetry, fluorescence and other factors can vary. Being able to extract the maximum yield from every rough diamond, in terms of both quantity and quality, can make all the difference — not just to the manufacturing firms’ profitability, but to their survival as well.
Sarine Technologies is a pioneer in AI-driven diamond-processing devices. Sarine’s systems scan rough diamonds, map inclusions, and generate multiple cutting and polishing options so manufacturers can maximize yield. The firm’s AI-based grading systems, meanwhile, bring heightened accuracy to assessing color, clarity and cut.
Training the algorithm — a key part of using AI effectively — entails feeding it a significant amount of data. “Having enough variation of the data so it represents the entire population of diamonds, and having that available to train systems, was a challenge when we started,” acknowledges Sarine CEO David Block. However, “that’s one of the advantages we have today with over 100 million diamonds going through our technologies annually.”
Another notable company using AI is the Lexus Group. Lexus India has integrated AI-based inclusion-charting into its Cutwise diamond-processing platform, enabling customers to get the maximum value from their stones. Lexus is also a leader in offering processing services and solutions for fancy-color and high-value diamonds.
Colored-stone computing
When it comes to colored gems, a stone’s origin and treatments have a significant impact on its price and its value appreciation over time. As such, there is an ongoing tug-of-war between lab gemologists and gem merchants, as the latter will often dispute the former’s reports on goods’ sources and treatments if those results don’t match the higher-value designations the merchants were expecting. Machine-learning algorithms could potentially ease those tensions by improving accuracy and reliability.
Gübelin Gem Lab, known for its legacy of gemstone authentication, is leveraging AI in its Gemtelligence system. This technology identifies intricate patterns and markers that can be difficult to spot through conventional methods, refining the process of determining origin and detecting treatment. In rigorous testing, Gemtelligence has achieved higher accuracy in predicting origin and pinpointing heat treatment than human experts have, even without microscopy, according to Gübelin. This lets the lab shorten the testing pipeline and offer better prices for some of its services.
But does this mean machine learning will replace human gemologists in labs?
Even with Gemtelligence significantly enhancing accuracy, human expertise remains essential, stresses Gübelin managing director Daniel Nyfeler. “Every result comes with a confidence percentage. If the system’s confidence about its conclusion drops below a set threshold — such as 98% for ruby heat treatment — the stone will undergo the traditional testing procedure, including two full descriptions by senior gemologists. And every single stone, regardless of the confidence level achieved by Gemtelligence, gets a quick visual plausibility check for a final verification.”
In addition, a gemologist’s role is vital in identifying gems from new deposits or those that have had new treatments, since AI models are only trained on historic data.
Most gem labs globally are using algorithms to evaluate stones’ chemical data, but when it comes to their mainstream operations, they are still testing the integration of AI. Gübelin may soon offer a “plug and play” option so other labs can connect to Gemtelligence’s database.
The artistic angle
One of the rapidly expanding segments within the larger jewelry universe is bespoke or custom commissions. How does a jewelry retailer or a manufacturer cater to this growing desire when resources are limited? Jewelry design, traditionally the domain of skilled artisans, is experiencing an AI-driven revolution as well.
Payal Shah, founder of fine jeweler L’Dezen, introduced an AI-powered text-to-design tool last year that lets users generate jewelry concepts instantly. With a simple text prompt, her system can create endless design variations to aid in creative expression. It can also save resources: Retailers don’t need to keep a large jewelry inventory when they can work with a manufacturer to offer bespoke pieces.
However, the rise of AI-generated designs raises ethical concerns. Shah acknowledges that AI relies on existing data, making originality a challenge. “It comes down to the human touch of integrity,” she states. “If someone prompts the system to generate a design in the style of Van Cleef & Arpels, that’s what it will produce.” Ethical AI usage, therefore, depends on designers ensuring their work remains innovative rather than derivative.
Pictures and product descriptions
Beyond manufacturing and grading, AI is also transforming the marketing sector.
Picup Media’s GemLightbox technology employs AI to capture high-quality images and videos of gemstones and jewelry. It can also generate accurate product descriptions within seconds — ideal for social media, e-commerce and cataloging.
While there are concerns that this could make jewelry descriptions too formulaic, Picup Media sales director Rocky Yu explains that “AI can’t replace human creativity, but it can support and enhance it. AI can handle the ‘boring’ parts of a description — like mentioning how a diamond is surrounded by colorful gemstones. But the story behind the piece — the craftsmanship, inspiration and emotion — that still needs to come from a human.”
As AI continues to evolve, its role in the diamond, gem and jewelry sector will only deepen. From precision-driven manufacturing to design and marketing, artificial intelligence is working behind the scenes to improve quality. While concerns about job displacement and creativity remain, those who embrace AI as a tool rather than a threat will find themselves ahead of the curve. The key to success is understanding how to harness its potential while ensuring that human expertise and ethics remain at the heart of the industry.
Main image: E-grading system Sarine Clear assesses a diamond. (Sarine Technologies)
