Know Your Honey
Fraud honey is a global threat to beekeeping. Proven traceability of authenticity and origin with AI and blockchain will change this.
Fake honey is a well know fact – the EU states 50% fraud and has identified 5 common methods, see below. The counterfeit affects beekeepers all over the world, large and small, and of course consumers.
The goal is to develop a new standard for authentic honey and Good Goods. Funding is sought through the Swedish Board of Agriculture, European Innovation Program.
Know Your Honey is a development of the BeeScanning technology, an interdisciplinary two-year project with several partners. The project is initiated and led by beekeeper Björn Lagerman, founder of BeeScanning. The partners are ATEA Sweden AB in IT and blockchain Technology, BeeLife (European Beekeeping Coordination, NGO), Biodlingsföretagarna (Swedish Commercial Beekeepers Association), European Professional Beekeepers Association, EPBA; E-sense AB, analysts in food chemistry, researchers from the Swedish University of Agricultural Sciences, SLU, with classic and modern honey analyses. The project also cooperates with Neurolearn AB in AI, a spinoff from Örebro University and for the app development with Gislen Software in India.
Cheating is visible on the shelves.
Image 1. What can be trusted in a swedish supermarket?
The EU will require all food to be traceable. New technology can ensure the traceability, authenticity and origin of a product. It is based on the combination of data from the app user, AI analysis of images, new analysis methods and securing data with blockchain technology.
Explosion in honey fraud
Honey is one of the most counterfeited foods in the world (note 1). Worth billions, it is easy to counterfeit but difficult to control. The value of honey globally is over SEK 90 billion in 2022 and is estimated to increase to SEK 260 billion in 2028 (note 18). Huge values are damaged for both consumers and producers by fraudulent handling of a unique raw material. Internationally, the need to combat the counterfeiting of honey is a high priority. Between 30-70% of the honey on the world market is more or less adulterated (note 11). A coordinated EU action in May 2023 states that about 50% is counterfeit (notes 12-13). Another good example is Asia: exports from East Asian countries have gone from 100 million tons in 2010 to 325 million tons in 2018. China exports twice as much ‘honey’ as it has bees for. Exports have tripled, but the number of colonies remains unchanged (Note 16).
Image 2. Honey? from China.
Despite great efforts to develop analytical methods to detect adulteration, this has failed and has not prevented further cheating (note 14). On alibaba.com you can buy “honey” for €1,4/kg, with a few extra cents you can get labels such as organic or Krav.
Image 3. Certified ”Non-EU-farming”, honey?!
The methods of transforming rice syrup with colors and flavors and even the addition of pollen are ahead of the analyses that can detect such cheating. Calls for stricter rules on origin labeling, increased controls and even import restrictions are becoming louder and more frequent. Unfortunately, this is more or less in vain as there yet are no tools to tackle this large-scale criminal fraud.
1. The small hobby beekeepers are not the problem. They have no reason to cheat; they are proud of their craft. You can usually trust their word that they are harvesting real honey. However, this is a matter of trust and there is no objective proof. However, hobby beekeepers are affected by cheating indirectly through lower valuation of the honey and distrust from some consumers (note 19).
2. Larger beekeepers, with a few hundred colonies or more, may have an economic interest in improving their performance by illicit means, especially if they are under pressure from both cheap foreign mixed honey as well as counterfeit competition. However, this is likely to be marginal in Sweden and in some parts of the western world. Beekeepers and associations tend to keep a fairly good eye on their colleagues locally. However, in some countries, such as the Middle East, there is a fairly widespread perception that cheating is legion, and it is therefore considered that comb honey provides some guarantee of authenticity. But there is no objective documentation to prove either the origin or the authenticity of the honey.
Image 4. Labeled frameless super with combhoney from FriBi HB
3. The way the honey market operates, cheating occurs mainly at the packer and retail level. Retailers are happy to do good business buying at low prices as long as the supplier can show the right certificate. Bottlers are satisfied with being able to buy for €1,5/kg, mix with real honey or sell directly to the trade for €10/kg. The consumer pays €15/kg. No objective evidence can be presented.
The strategy consists of four parts: 1. Labeling supers, 2. An app with manually and automatically collected data, 3. Image analysis with AI, and 4. Blockchain technology.
The project offers a unique solution based on verifying the raw commodity at the time of harvest with AI analysis of images of the harvest before it ends up in a container. In the case of honey, it needs to be linked to harvesting from a specific bee colony in an identified hive in a precise location (see Figures 4 and 5 for an example of labeling honey supers). Beekeeping in the world has so far lacked possibilities for objective identification of frames and boxes. This is related to past difficulties in documenting, verifying and recording labeled units. The labeling itself can be made very simple without costs.
Image 5. Weighed and labeled super.
All events are linked in a chain that can not be corrupted backwards. Primary data is created by the beekeeper using the app. When photographs are taken of the hive and colony, data is collected containing; colony quality and health status (BeeScanning’s patent), identity, time, location, number of colonies, grower’s data, local conditions, meteorological data, flowering, pollen occurrence, harvest data. Primary data can be correlated with secondary data from growers in the same region and, if necessary, various classical and modern honey analyses.
Image 6. Events from producer to consumer generating digital honey profile.(Click to view)
Validation of the authenticity and origin of honey is thus not based solely on analyses of honey, but the analyses can constitute so-called proof points in a chain of transactions. The project will determine which analyses will be integrated, and the possibility of authorizing analysis providers, in order to maintain the quality of the input data in the system. The analyses can constitute “fingerprints” for validation and include: pollen content, chemical analyses of sugars, fragrances, other substances such as environmental toxins, pesticide residues, microbiological analyses of harmful and beneficial microbes, and last but not least DNA analyses to map the honeybee’s macro and micro environment, and possibly images from MRI analyses.
Image 7. Example of a close-up of a honey sample at 4500 x magnification. The image shows pollen, various microbes and other substances. With simple microscopes used for classical pollen identification it is not easy to detect microbes, but this image shows that there is a diversity of organisms and substances that we do not know much about. Modern methods that detect both microbes and DNA can help characterize a honey sample in much more detail than classical chemical methods can do. Image taken at the Umeå Center for Electron Microscopy by Natuschka Lee.
The entire chain is secured and validated via blockchain technology. With secured data, the honey industry can demand the individual steps of a complete solution. Furthermore, this can give the beekeeper constructive feedback on the quality of his honey and what measures may need to be taken to improve the quality of the production. This is a unique approach that can serve as a model in the world for honey and other raw materials.
Disclosure of fraudulent practices
Fraudulent practices defined by the EU (note 13):
1. Use of sugar syrup (from rice or corn) to adulterate honey and reduce its price, both in non-EU countries and on EU territory.
Revealed by: The quantity cannot be linked to a defined origin. Missing, inaccurate, discrepant producer data, hive ID data, treasure box, GPS image evidence, time of harvest and AI analysis of harvest image. Missing data from extraction. Missing secondary data on crops and harvest in the region, missing reference sample.
2. Analysis in accredited laboratories to adjust mixtures of honey and sugar to avoid detection by customers and official authorities before import.
Revealed according to 1 above complemented by DNA, pollen, microbes and NMR analysis with reference library.
3. Use of additives and colorants to falsify the true botanical source of the honey.
Revealed according to 1 and 2 above.
4. Masking the true geographical origin of the honey by falsifying traceability information and by removing pollen, or adding foreign pollen.
Disclosed by blockchains securing the data under 1 and 2.
5. Feeding of bee colonies with sugar during nectar flow
Revealed by wrong sugars and their relationship. Wrong pollen content. Revealed by unreasonableness of harvest in relation to secondary data such as available nectar sources, comparable data regional production, comparable data history. Revealed by audit of financial accounts.
The fraudulent practice of feeding bee colonies with sugar solution falls under a combination of the points above and would thus be revealed by one or more of the points of evidence in the events building up the block chain. If an unusual amount of honey is produced from a specific area at an unusual time of the season, the system may raise an alarm and warrant further investigation. Nectar production in an area is limited by several factors, the plausibility of which can be assessed externally based on secondary data. Examples of such data include meteorology, soil fertility, crops grown, density of bee colonies, history and satellite monitoring via remote sensing of environmental conditions. In case of discrepancies or inconsistencies between data, the technology can warn and point out the deficiencies and reject an approved digital honey profile.
China’s (note 15) production of twice as much fake honey as there are bee colonies would be made impossible as described above.
The product is an activated QR code on the packaging that allows the consumer to scan and verify the origin and authenticity. This connects the customer with the producer’s images and information from the different stages of production. The producer can choose how deeply the customer can look into all the details of the production.
Value: The beekeeper and honey extractor can be better paid if they can gain the customer’s trust that the product is genuine and the cheapest fraudulent products are limited.
Voluntary: The system will be voluntary to use. When consumers value the ability to see the origin of the product, demand is created.
Competition: For the producer, the beekeeper, traceability becomes a competitive tool against those who cannot provide objective traceability.
Potential Economic Value Added Through Price Premiums (note 9)
We envisage a model based on a fixed and a variable component. The fixed part consists of an annual subscription and varies with the number of hives. The variable part varies between 0.5%-5% of the product value. The impact of traceability on the value of the honey can be set at an increase in value of 10-30% at each stage of the transactions.
The regulatory framework
EU requirements to develop traceability for all foods will affect all stakeholders. Everyone will have to find tools to comply. Consumers may demand it, when technology makes it possible to know the origin of a product.
AI analysis of raw materials has the potential to become a very powerful quality tool. The technology will fundamentally affect food production. To the benefit of safety, honest practices and good quality.
Call for proposals/participation
We welcome participation to suggest ways to prevent practices that try to circumvent the system. We seek collaboration with beekeepers who can provide honey supers for photography or images of these for training the AI.
Image 8. Images of supers for training AI.
We would also like to collaborate with testers when the first versions become available in 2024. A final product is planned for the end of 2025.
Björn Lagerman, BeeScanning, firstname.lastname@example.org 070-5603893
Natuschka Lee, Swedish University of Agricultural Sciences and Swedish Collegium for Advanced Study, email@example.com 070-3751213
9. https://jisar.org/2022-15/n1/JIARv15n1p24.pdf sid 24-30
10. https://www.fao.org/3/cb5353en/cb5353en.pdf sid 146
19. Land Lantbruk nr 35 25 augusti 2023, sid 6-10.
All images BeeScanning except #2 (note 17) and #6 Natuschka Lee.