How Optics Detects NFT Fraud with Artificial Intelligence and Machine Learning – The New Stack
The NFT space has ongoing problems with fraud, including through bad actors who lift art from one project and use it in another project—a process often referred to as “copyminting.” They are derivative projects that have a bit too many similarities to the original project to be considered anything other than a ripoff. Although most of these duplicate projects have very little sales volume compared to the original, they can damage the underlying brand, contribute to the general distrust of the NFT space, or trick less savvy buyers into spending money on something that is the jpg equivalent to a street shilling selling fake Rolex watches.
To help combat this fraud, a few companies are emerging that specialize in fraud detection in NFTs. They tend to leverage blockchain data to find out which project came first and use some image recognition to find metadata matches. One of these solutions is Optic, which uses artificial intelligence and machine learning to analyze the images associated with an NFT, helping NFT marketplaces and minting platforms to catch copies and protect both creators and buyers.
In a recent press release, Optic indicated that they are the copymint detection partner for OpenSea, arguably the best-known NFT marketplace, which introduced this feature in May 2022. Optic currently processes around 2 TB of NFT metadata per day, from millions of newly minted NFTs. They analyze for both exact copies and fuzzy matches – including flips, rotations, resizing and color changes.
AI and machine learning powered API
The AI and machine learning behind Optic leverages NVIDIA Triton Inference Server, Milvus, PyTorch’s machine learning framework, MLflow, Amazon SageMaker, Lightning and open source ML version control via DVC.
The New Stack has a guide on deploying Inference Server, as well as a number of articles on SageMaker, if you’re interested in exploring AI for your own projects. Given the potential complexity of making all these technologies work together, if you’re considering a similar solution for analyzing NFT metadata, it might be easier to implement the Optics API than trying to train an in-house solution.
Optikk supports image analysis via either a batch processing API or a real-time API. I asked Andrey Doronichev, CEO and founder of Optic, on the difference between the two. He said: “In batch processing mode, a marketplace makes its catalog available to Optic using a data store (eg S3 or Snowflake), Optic reads new data periodically, checks each collection and returns match results back to the marketplace. real-time mode calls the marketplace Optic API, provides token metadata, and Optic returns match results to determine whether it is original or a copy coin, as well as which is the original and what is the match result.”
You can see a sample comparison along with the Optic score in the screenshot below.
One of the areas where I have seen copying to be a frequent problem is in cases where the original release may be on Ethereum, with a copy of the project launched on Solana or Polygon. I asked Doronichev how Optic addresses this. “Cross-chain copyminting is a real problem,” he said, “which is why we currently match tokens across Ethereum, Polygon and Solana chains. We will prioritize which chains we support next, based on NFT -use.”
While the Optic API is currently private, they plan to make a public API available before the end of 2022. In the meantime, if you’re building a marketplace for NFTs or just want to make sure users aren’t trying to exploit other people’s intellectual property property, you can contact Optic to request access to their private API or join their Discord to engage with the team. There is also a white paper on the Optic protocol coming soon.
Lead image via Shutterstock.
