Humans and computers to fight NFT plagiarism on OpenSea
Non-fungible token (NFT) marketplace OpenSea is attempting to tackle plagiarism on its platform. A new “two-pronged” approach will aim to root out and remove copies of authentic NFTs and changes to its blue tick verified badge system will seek to enhance customer confidence.
According to a blog post, the anti-plagiarism system will rely on the automatic detection of so-called “copymints” which will be supported by human reviewers to keep the algorithmic detection system in check.
As noted by The Verge, last year, NFT collectors using the platform were tricked into buying flipped images of the popular BAYC collection. The largest NFT marketplace was also forced to remove its delay pay minting service after an influx of copymints of existing collections.
At that point, 80% of NFTs removed by the platform were minted through its lazy minting service which postponed gas payment by keeping the digital asset off-chain until the first purchase of the NFT.
Copyminted NFTs are opportunistic cash grabs riding the hype wave of popular collections. A copyminter will literally right-click, save the files and mint the jacked JPEGs that are then flogged under the guise of an authentic NFT.
According to the blog, OpenSea’s anti-plagiarism system will scour the platform’s many collections for flipped and rotated images as well as other currently undescribed permutations. The lack of detail here could be a tactic to keep information away from scammers but also it could allude to the new system’s present rudimentary nature.
That’s where the human detectors come in who will likely verify images flagged by the detection system and help to train the system to work more independently.
“With this system, our long-term goal is two-pronged: first, with help from our community, to eliminate all existing copymints on OpenSea; and second, to help prevent new copymints from appearing in the first place,” it said
“We’ve already started the process of delisting identified copymint collections, and we’ll scale up our removal process in phases over the next several weeks.”
It works for Twitter, why not OpenSea
OpenSea is also making changes to the way it verifies accounts. Initially, accounts with an NFT stack worth 100 ETH ($196,000) or more will be invited to apply for verification. OpenSea said it also plans to offer the blue checkmark to people who are active and have a “presence” on Twitter and Discord.
“This is just the start – we are committed to a future where any authentic creator’s account can be verified, while keeping scammers out of the system,” it said.
Read more: DeviantArt enlists AI to police NFT markets for stolen art
Collections will also receive a blue checkmark if they are owned by a verified account or if there’s a lot of hype around it and a substantial amount of sales. For example, a new NFT collection put up for sale by the verified BAYC account will come with a blue badge so potential customers know it’s legit.
OpenSea said that any already ‘verified’ accounts and collections will not lose their blue tick and new requests will be responded to within seven days.
“We consider these programs and products collectively as OpenSea’s lighthouse – helping protect creators and collectors and ensuring our community can navigate the world of NFTs confidently.”
“Together, we believe these changes simultaneously improve trust in the NFT ecosystem by elevating authentic content and removing plagiarized works,” OpenSea said.
Indeed, OpenSea is merely catching up with other areas of the digital art economy. In August last year, the online art portfolio platform DevianArt upgraded its artificial intelligence (AI) image recognition system to find stolen tokenized art on marketplaces like OpenSea.
DeviantArt Protect scans all uploaded art and cross-references with “public blockchain events” like Ethereum based NFTs to spot stolen art.
For more informed news, follow us on Twitter and Google News or listen to our investigative podcast Innovated: Blockchain City.