Top Web3 Fraud Detection Start-Ups to Watch

Scammers and hackers drained over $504M from web3 protocols in just the third quarter of 2022. According to a report by the blockchain audit company Certik, exit scams and flash loan attacks were the two of the most common fraud types. Whether through these strategies or by creating synthetic identities, account takeovers, IP infringement, or paying with stolen payment instruments, fraudulent activities pose a severe threat to all Web3 participants who are in the space with good intentions. 

Web3 fraud detection start-ups combat malicious actors by merging multiple technologies that leverage machine learning and artificial intelligence – like behavior biometrics, AI graph, and neural networks- and tailor them to web3-specific situations. 

Four Web3 start-ups offering fraud detection solutions

Sardine detects scammers via behavior biometrics 

Start-ups can detect fraud by behavior biometrics; among them is Sardine. This technology uses machine learning to measure and analyze patterns in human activity. It helps differentiate legitimate users and fraudulent actors by detecting anomalies in behavior. 

Sardine applies a consortium approach to building supervised machine learning models and training them about fraud patterns. The blockchain company also developed a risk SDK that can detect suspicious devices and sessions. 

These technologies can be applied to fight against many fraud types in real-time, including those based on stolen credit cards during NFT checkouts. For this purpose, Sardine provides a payment option with integrated fraud and compliance technology. This solution permits purchasing NFTs for legitimate users while preventing bots from doing so. For instance, in a partnership with Autograph, the NFT purchase conversion rate of legitimate users could climb to 94% because Sardine payment was able to identify 98% of the users as not fraudulent. 

Cylynx powers fraud detection by Graph AI

Another way to tackle fraud is to include graph data in the detection process. Graph AI enhances machine learning algorithms to infer context and gain improved insight into relationships. Consequently, companies can unveil hidden connections and spot signs of fraudulent activities. 

Powered by this technology, Cylynx develops three solutions for its clients to deploy in situations related to web3 security. For one, there’s the Motif, a no-code visual interface. Network visualization is a central topic in fraud detection. And Motif speeds up the process by simplifying the integration of graph data into business intelligence.

Another product by the start-up is Transaction Monitoring which helps perform watchlist checks against international money laundering blacklists, spot malicious activities through behavior analysis and hold transactions in real-time if the algorithm finds them suspicious. 

There’s also the Crypto News API that scraps news websites and generates risk scorings based on a sentiment analysis model to enable tracking crypto risks. 

AnChain supports developers in building compliant solutions via a patented risk engine 

The technology of AnChain incorporates behavior patterns into a machine learning engine to estimate risk scores and update them based on real-time transaction behaviors.

Blockchain Ecosystem Intelligence API (BEI API) – developed and patented by the start-up – can screen and onboard wallets to detect fraudulent transactions. It connects wallet addresses to real-world entities and identifies important players – including those involved with scams – within a crypto ecosystem. 

Optic detects copymints with neural networks tailored for NFT infringements 

The fraud detection start-up Optic develops an AI engine powered by a neural net for NFT content recognition. This helps understand if the token is original by comparing it with authentic collections. 

OpenSea teamed up with Optic to fight copymints, for example. The start-up processes 2TB of new NFTs added to the world’s largest NFT marketplace daily to authenticate the assets. Its technology helps identify exact copymints, as well as alterations such as rotations or color modifications, within 400ms and with 99.9% precision. The same technology can be leveraged in other NFT marketplaces as well. 

The start-up will soon provide it to creators. So that they can verify their work, track copies across multiple NFT marketplaces thanks to scam reports and submit takedown requests.