AI Resilience Report forGem and Diamond Workers

If you're worried about robots taking over gem work, here's the honest picture: AI is already deeply involved in diamond grading, but mostly as a smart assistant working alongside skilled humans. Since around 2016, Sarine developed an automated color and clarity system, with GIA partnering with IBM the following year in 2017, because these organizations had the huge datasets needed for machine learning research ^[1]. Today, DeBeers uses a "Falcon" system for color and an "Eagle" system that captures high-resolution 3D images of inclusions and automatically compares them to a database to assign a clarity grade.

Gübelin's "Gemtelligence" examines analysis results for patterns that may elude human gemologists, and in testing it has surpassed human gemologists at detecting origin or heat treatment — though any result below 98% confidence is automatically rechecked by senior gemologists. A brand-new Swiss lab called SIG launched in April 2026 explicitly marketing "AI-assisted" colored-stone reports ^[2] as a co-pilot, with senior gemologists making the final call. Industry leaders are blunt: at the 2025 CIBJO Congress, Sarine's CEO said AI "will replace the repetitious, tedious jobs that are done in the lab" ^[3], while emphasizing the change is gradual.

Adoption is moving fast in big labs but slowly elsewhere, for a mix of reasons. On the speed-up side, diamonds are perfect for AI: they exist in huge, standardized volumes, and automated color grading offers consistency, repeatability, and round-the-clock operation that human eyes can't match. Economic pressure is also intense — the global diamond business sank deeper in 2025 as weak demand and cheap lab-grown alternatives shook the industry, with De Beers announcing plans to cut more than 1,000 jobs ^[4], pushing labs to slash grading costs.

On the slow-down side, small labs worry AI will take business away from those unable to invest in their own systems ^[5], and jewelry buyers crave human emotion and storytelling that algorithms can't fake. As one National Jeweler columnist reminds readers, AI is best treated as a tool jewelers can practically leverage ^[6] rather than a full replacement. The encouraging takeaway: gemologists who embrace AI as a tool will be better positioned, because humans excel at discovering what remains unknown — new materials, new treatments, and new synthetics that AI hasn't been trained on yet.

Customer advising, artistry, and judgment calls remain very human jobs.