Understanding Origintrail (TRAC): Building a trusted knowledge foundation for AI

The aim is to help organizations, communities, and individuals combat misinformation in the age of AI. At the core of the OriginTrail ecosystem lies the Decentralized Knowledge Graph (DKG), which combines blockchain and knowledge graph technologies to bring a global network of connected and verifiable knowledge that AI systems can rely on to life. The core elements of the DKG are interconnected Knowledge Assets (KA), each discoverable, verifiable in terms of provenance, and underpinned by an NFT that affirms ownership.

OriginTrail DKG is structured into two main layers:

• the blockchain layer – ensures security and consensus through a multi-chain design, and facilitates trusted knowledge exchange across Web3 communities. The latest iteration of OriginTrail DKG is running on the purpose-built NeuroWeb blockchain on the Polkadot and Gnosis blockchain
• the knowledge graph layer – a decentralized, permissionless network of nodes that host and ensure accessibility to the global knowledge network.

Fueling the DKG is the trace token (TRAC), used to publish KAs, expanding the global network of knowledge, and as collateral on DKG nodes, ensuring that node operators and TRAC holders (via delegated staking) maintain a vested interest in the network’s stability and security.

Currently, the OriginTrail ecosystem is in the Metcalfe evolutionary phase, named after Ethernet co-founder and Internet pioneer Dr. Bob Metcalfe, who advises OriginTrail’s core development team, Trace Labs. This phase focuses on scaling the DKG's capabilities towards the ambitious goal of creating the world’s largest Verifiable Internet for AI, reaching over 100 billion KAs, and introduces several new concepts to facilitate that growth such as knowledge mining and signaling, decentralized Retrieval Augmented Generation (dRAG), paranets and Initial Paranet Offerings (IPOs), and autonomous knowledge graph growth. More details on these concepts are available in the OriginTrail Whitepaper 3.0 on the Verifiable Internet for AI.

Meanwhile, the OriginTrail DKG is already powering trusted AI solutions, including those used by Fortune 500 companies, across industries such as FMCG, life sciences and healthcare, transportation, compliance, global supply chains, and several others.

How was OriginTrail developed?

OriginTrail founders Žiga Drev, Tomaž Levak, and Branimir Rakić set out to empower organizations to exchange data in a trusted and transparent manner.

Initially focusing on food supply chains more than a decade ago, their vision ultimately led to the creation of the OriginTrail DKG and the deployment of DKG-based solutions across several industries. As AI technology advanced , it became evident that addressing AI-related misinformation - issues like hallucinations, bias, ownership, and general verifiability - required a robust foundation of verifiable knowledge. The DKG emerged as this critical component, now underpinning trusted AI applications for global enterprises and organizations including the British Standards Institution (BSI), Supplier Compliance Audit Network (SCAN), Swiss Federal Railways (SBB), Perutnina Ptuj (the largest poultry producer in SE Europe), the World Federation of Hemophilia (WFH), and several projects under the EU and UK governments.

How does OriginTrail work?

OriginTrail DKG empowers organizations, communities, and individuals to transform their data into KAs, making their knowledge discoverable, verifiable, and usable by AI solutions. As KA publishers, they compensate DKG nodes with TRAC tokens to maintain this expansive network of knowledge.

Within the DKG, KAs are organized into paranets, which are focused clusters of KAs, ontologies, and AI services centered around specific topics. This organization facilitates the enhanced aggregation and extraction of knowledge for AI solutions. Further, paranets can be launched through the Initial Paranet Offering (IPO) mechanism, adding an incentivization component whereby knowledge miners are rewarded for enriching a paranet with new knowledge (more details here) . This allows organizations, communities, and individuals to crowdsource valuable knowledge onto the DKG so it can be used in their trusted AI solutions.

The mechanism powering AI solutions on OriginTrail DKG builds on Retrieval Augmented Generation (RAG), a technique for enhancing the accuracy and reliability of Generative AI models with facts fetched from external sources. Decentralized Retrieval Augmented Generation (dRAG) enhances the technique by extracting verifiable knowledge from OriginTrail DKG, using Knowledge Graph techniques for more precise information retrieval. It’s compatible with any AI model (e.g. OpenAI GPT, Google Gemini, X Grok, Meta Llama, etc.), providing flexibility and choice to meet various use case needs. AI solutions built on the DKG, such as chatbots, use dRAG to provide users with relevant answers and a way to verify the sources used by the AI model. A great example of such a solution is the one used by Perutnina Ptuj, the largest poultry product in SE Europe, which enables consumers to interact with the brand and inquire about the origins of their meat purchases - you can access it here.

How is the TRAC token used? 

The Trace token (TRAC) is used to power the OriginTrail DKG. It was launched as an ERC-20 token on the Ethereum blockchain in 2018 and is used for:

  

• Publishing and updating Knowledge Assets
• Collateral on DKG nodes to ensure that nodes are performing the required network functions. This increases their chance of being selected for compensation for holding selected Knowledge Assets
• Transferring of value

The TRAC token is increasingly being used by network node operators and Knowledge Asset publishers alike. More live statistics of the DKG and TRAC token use can be found at the community-operated OThub.io website.

Token distribution

  

There is a fixed maximum supply of 500 million TRAC tokens. Half of these were sold in the presale/crowd sale in January 2018, when all tokens were minted simultaneously. There is no token inflation since no further issuance of tokens is planned.

Of the remaining tokens, the team reserved 20% of the total supply for future project development, 18% was allocated to the founders and pre-ICO contributors and investors, 5% was dedicated to ensuring the token’s liquidity in crypto markets, 5% was promised to the development team and advisors, and 2% was reserved for rewarding those who facilitated the token sale. 

Conclusion

 

Today, we’re facing a reality where AI-generated content is increasing exponentially, causing significant problems with misinformation. OriginTrail is pioneering a Verifiable Internet to enhance the trust in AI and maximize its potential by addressing critical challenges such as hallucinations, bias, and the verifiability of AI-generated content.

OriginTrail's Decentralized Knowledge Graph (DKG), powered by the TRAC token, harvests the power of knowledge graphs and blockchain to establish a robust foundation for verifiable and interconnected knowledge that any AI system can utilize to enhance trust.

With innovative frameworks like decentralized Retrieval Augmented Generation (dRAG), OriginTrail is already powering AI solutions across diverse industries, including compliance, global supply chains, international trade, and real-world assets (RWAs). Trusted by prominent organizations such as BSI, SCAN, Swiss Federal Railways, and many others, OriginTrail is setting a new standard for AI, rooted in Verifiable Internet.