top of page
  • GlitteratiGuestEdit



In the past, people relied on weather forecasts issued by old, experienced men who carefully studied the drifting clouds, possibly complemented by the flying behaviour of bees and birds, and the pain level in their knee joints. Their credibility was based on their decades of experience and their aura as wise men.

Nowadays, we rely on weather forecasts produced by professional institutions, on the basis of a global network of weather stations collecting large amounts of data, which are fed into digital models based on empirically-derived algorithms.

Since many decades, the gemstone industry, specifically the coloured gemstone trade, largely relies on gem lab reports issued by experienced gemmologists, who carefully look into microscopes, possibly complemented by some spectra and chemical data, and expressing an ultimate verdict on a stone. Their credibility is based on their decades of experience and their aura as a gem expert.

This – admittedly exaggerated – comparison triggers the question; why does an entire industry come to accept spending serious money while relying on lab reports based on a concept, which is considered outdated in other domains of our lives.

The history of gemmology is the history of famous gemmologists. We remember names such as Robert Webster, George F. Kunz, Robert M. Shipley, G. Robert Crowningshield, Richard T. Liddicoat, Alan Hodgkinson, or Edward J. Gübelin, succeeded by more recent, but already acclaimed individuals that populate, and often lead, today’s gem labs. By virtue of their work, their knowledge and passion, these experts have achieved recognition across our entire industry, and have shaped what we today understand as gemmology.

Each of them has studied tens of thousands of gemstones and built up an expertise making them true gurus in their field, and inspiring generations of young gemmologists. Based on this vast level of experience, their judgement on the nature, authenticity, and sometimes also on the origin of a gemstone was, or still is, considered axiomatic.


However, that image of the quasi-infallible gem expert has started to crumble in the last decades. Lab reports with conflicting or undeterminable origins are known from all major gem labs. So are cases of cultured pearls getting reports stating natural provenance and vice versa. Treated green diamonds get lab reports stating natural colour, and some types of treatments remain undeterminable, or even go completely unnoticed by gem labs. Aside of massive losses (and also massive gains) in value for the owner of these stones, such cases also caused some disillusion in the trade. Nevertheless, the relation between the gem trade and gem labs remains symbiotic, exemplified by the branding of origins and trade colours, which, despite the obvious limitations, keeps fueling the trade of coloured gemstones. Proper remedy to overcome these limitations would be to address verifiable transparency in the supply chain of gemstones. The work of gem labs, including the determination of country of origin, is forensic by nature, and remain imperfect. The ever louder demand for lab harmonization – whatever that means – is symptomatic for a somewhat frustrated industry.

The reason for such gem lab errors are manifold. One is the increasing complexity of the matter – new types of synthetics, sophisticated and hard to detect treatments (such as low-temperature heating applied to corundum), new origins producing gemstones which were previously considered as unique from one location only (such as some types of blue sapphires from Madagascar with properties confusingly similar to Kashmiri sapphires). To keep abreast with such developments, gem labs were forced to upgrade their arsenal of analytical tools. The microscope – the weapon of choice of the traditional gemmologist – today merely remains as one source of insight, alongside other more quantitative methods. To ensure that these complex analytical instruments deliver meaningful and reliable data, specialists with specific skills get assigned with these new tasks. In addition, the expansion of analytical weaponry has triggered an explosion of the amount of data collected on a single stone, for instance with the application of mass spectroscopy and computer tomography. Human brains, however brilliant, can only consistently process a small number of observations at once. The larger the amount of data, the less consistent the processing is by the human brain. Furthermore, well described cognitive mechanisms, such as confirmation bias, anchoring or the primacy effect, influence human observation and affect the interpretation. Such effects potentially reduce the repeatability and hence the consistency of the final conclusion. But consistency of results is of foremost importance for gem lab clients.

Other risks are looming: When confronted with complex systems, the human brain tends to handle it by reducing complexity. The desperate search for so-called “diagnostic” features in a gemstone, or the evaluation of chemical data by means of binary diagrams are typical symptoms of the misleading attempt to simplify highly multivariable systems.

Consistency of results requires standardised procedures of data gathering, and a transparent, repeatable method of evaluating the data. Human experts often are limited in both these tasks. First, they seem to favour a high degree of methodical freedom when assessing a stone. Secondly, the processes of observation and description are often intimately interwoven with their interpretation and conclusion. This conflicts with one of the underlying principles of scientific work to strictly separate observation, interpretation and conclusion. Consequently, many experts struggle to explain plausibly and in detail how they come to a specific conclusion.

From our own experience with many gem experts, we came to realise that analysing and understanding the mechanism of an expert’s decision-making is a challenge, and properly disentangling observation from interpretation is often impossible. This makes it difficult for many experts to share their knowledge, and can add an esoteric – in the true sense of the word - aura to their work. That might, maybe unintentionally, strengthen their status as an expert, while the intransparency of their evaluation process might nurture a nimbus of untouchability.


But gem labs led by such guru-like experts face several problems. First, these labs struggle to grow, because the final say is with one person only, and this one person has a finite capacity of seeing stones. Second, the eventual succession to another lab leader, be it an expert or not, is not smooth, and is usually accompanied by shifts, drifts and spikes in the interpretation, and hence also in the results, meaning the long-term reproducibility / repeatability of gemmological conclusions is lowered.

Gem labs that want to grow, and/or ensure long-term consistency for their clients have to say farewell to the guru-concept and embrace a new working model, adopting the more systematic approach of laboratories in other industries. We propose a model similar to modern weather forecasting based on multidisciplinary standardised data gathering and automated data interpretation. The data gathering is distributed to multiple specialists, meaning that the respective knowledge is dispersed over differently skilled and educated people. Most of the gathered data is structured and computable, i.e. knowledge can be stored in databases and processed by software, rather than only in the head of individuals. Data interpretation is taken care of by software running empirically-derived algorithms. They replace the intuition of the guru.

In recent years, the capabilities of computer models and algorithms to read and process variable types of data simultaneously has grown exponentially. It goes beyond the crunching of structured, quantitative data, extending into auto-evaluation of imagery and other less structured data. Interpretation of such heterogeneous sets of data by powerful pattern-recognition software allows new insights, independent of existing wisdom. Machine-learning and Artificial Intelligence (AI) methods also penetrate the field of gemmology, and redefine the gem lab work.

Today’s gemmologists need to adopt a new role of a specialist working in a multidisciplinary team on the same level as other specialists such as spectroscopists and chemists. They have to follow strict testing protocols, standardized methods and a system of checks and balances. This new model conflicts with the self-perception and image of the guru, who typically makesdecisions alone and in a partially opaque way.

In most gem labs, however, the formal or informal leader, be it called Director, President, CEO or Chief Gemmologist, is still a representative of the old model, i.e. a guru-type expert. They often hold a lot of decision power on the result of a specific stone and hence on the gem lab report. Quantitative data gets collected and considered, but the weight which is assigned to the different observations is often done in a non-systematic, sometimes arbitrary way. Such a system, often not even following a true 4-eye principle, can be assumed to be more prone to inconsistency. These experts also have the tendency to rely on their past experience, and struggle to integrate new evidence in a correctly weighted and unbiased manner to their existing knowledge.


We believe that model-based results are more repeatable and consistent over time than the expert’s results. This adds another major argument in favour of the new method as it allows for more transparency in the way a gem lab does its work. In meteorology; temperature, humidity, wind speed data and the like are publicly available. Gem labs, however, are only starting to think about giving access to the raw data they collect on clients’ stones. Giving insight into the raw data would force labs to explain how they work. In return, they would get challenged, and would need to explain or improve their methods. Such transparency would ultimately result in a more robust and resilient system.

So. Is this the end of the old-school gemmologist? Not exactly. They will not become obsolete, nor willthey vanish into anonymity. There is a type of knowledge beyond the one contained in databases and sophisticated algorithms. A machine cannot instill the enthusiasm and passion for gemstones as the human experts can. Their gut feeling to spot the new, unknown and out-of-the-ordinary is still unmatched by any algorithm. And it remains advisable to test the plausibility of machine-generated results with the critical eye of the human expert. The judgment and wisdom of the experienced gemmologists is still an invaluable asset needed in every gem lab, but applied differently. Integrated in a team, gemmologists can shift to other, more rewarding tasks, such as studying samples from new sources, maintain the databases that underlay the algorithms, oversee the quality of the results and share their knowledge with younger generations of gemmologists. And by spending more time on research and development, they can achieve visibility and personal fulfillment.

To bring gem lab work to a higher level of professionalism, it is required to disrupt the prevalent system in most of today’s gem labs and overcome the idolised guru-gemmologist mindset. Redefining their roles and specifically their influence on individual stones is not a sign of their redundancy, but of professionalism.

We assume that Dr. Eduard Gübelin, our idol, guru, and still our moral guide today, would agree.

Follow Gübelin Gem Lab on Instagram @gubelingemmology

#IndustryNews #GubelinGemmology #GemGurus #DanielNyfeler #TheGlitterati

bottom of page