[RR-8] AI-Enhanced Safety Feature (SCOUR)



We’re going to increase trust and safety across the Rarible platform and NFT ecosystem utilizing a combination of knowledge graphs, predictive analytics and deep learning.


There are great use cases for predictive analytics and AI models to enhance safety, scalability and trust in the Rarible platform.

The most impactful way to strengthen the platform is with wash-trade detection and flagging models. We are building a detailed knowledge graph of the complete Rarible transaction history including transfers, wallet addresses, and reward token distributions. With this dataset, we can train models to detect, flag and grade malicious or suspicious traders. This will drastically increase platform trust by detecting and stopping wash trading in a manner that is fast, reliable and scalable.

We call this malicious trader detection model, SCOUR .

Currently the Rarible team can receive up to a thousand reports/support requests per day. We’ve hit a bottleneck in the human capital which provides our platform support. So let’s relieve that pressure off the Rarible support team so that they can get back to helping people.

Economic Impact

In our analysis of 1/3 of confirmed malicious trader wallets; 2.95 million $RARI tokens have been minted to addresses which were involved with malicious trading activity. This represents 14.42% percent of the $RARI token distributions and $17.24 million in value at current prices.

14.42 % of current $RARI rewards represents $324,450 per week in platform reward losses to malicious activity.

Proposal Milestones

  1. Receive wash trader data directly from Rarible support. :heavy_check_mark:
  2. Deploy data as a knowledge graph :heavy_check_mark:
  3. Exploratory data analysis :heavy_check_mark:
  4. Data wrangling, cleaning and feature engineering :x:
  5. Preprocessing, data pipeline and proof of concept model :x:
  6. Minimum viable product deployment with API :x:
  7. Integration into existing Rarible support infrastructure :x:

Our Team

I am a creative futurist, data and AI tinkerer, and an Ocean Protocol Ambassador. I am from Bellingham, WA in the United States.

The Bitscrunch team provides big data, analytics, AI-solutions and devops. They hail from Munich Germany. Bitscrunch is a rapidly growing team with skills in computer science, devops engineering, data science, bioinformatics and their CEO Vijay Pravin was featured as a TedTalks Data Storyteller speaker.

Our team has been ideating and developing NFT marketplace platform safety features since October of last year. We have put a lot of thought and energy into this solution space.

Here is the image forgery and provenance detection model that we built for the Untitled NFT hackathon last year:

( https://kylewebster.medium.com/n-safety-forgery-detection-model-untitled-nft-hackathon-submission-11fb9c2b01b1 )

You can test out our image forgery detection model right here:

( live demo https://nsafety.bitscrunch.com/ )

Funding Request

We are requesting 2000 $RARI upfront and 2000 $RARI to be distributed according to milestone completions. The second half of the funds will be deposited into a Gnosis Multisig Safe — with signatory members from all the respective teams.

Four signatories will be on the Gnosis safe, 2 from our team and 2 from the Rarible team and/or community.

A portion of these funds (20%) will be used for expenses and infrastructure. That includes transaction fees and contract deployment fees as well as the procurement of a server that can continuously handle marketplace support prediction requests, data aggregation and model training.


We’ve developed a number of relationships with Rarible team members including Alex Salnikov, Eric Arsenault and Ethan van Ballegooyen. We will be working hand in hand for milestones eight through ten above. It’s very important that the finished product adheres to Rarible’s technical specifications.

Again funding will be distributed according to the completion of the above objective milestones.

Future and Goals

We’re very interested to continue deploying AI-enhanced safety features in the NFT marketplace space — like the image forgery and provenance detection.

Our goal for this MVP is to reduce $RARI tokens rewarded to malicious actors by a minimum of 15.4%


What is the timeline for the deliverables?


3 months Eric, will discuss over Tele about this with a detailed plan

To do it over millions of transactions on a live environment is something different from just an ML algorithm. Building an algorithm is different from taking it to Production. Unfortunately only 1% of ML algorithms are moving into production.

1 Like

With the amount of Wash trading happening which is about 17 million , the concern shouldn’t be on the funds requested. It should be about the deliverables and how effectively we can prevent this.


Neyadi do you have a LinkedIN? We’re certainly interested to collaborate with other data scientists on this initiative.


I think this is pretty spot on.

Tuning the model - is a small part of the issue. Most of the work is in the preparation, arrangement and engineering of the data - and the rest of it is the application programming interface and integrating it with the support team’s existing workflow.

Exactly my point as well Naresh