Technological progress is transforming the art authentication process. Techniques such as scientific imaging have allowed us to peer beneath the surface of paintings; advanced fluorescent lighting has allowed us to identify specific pigments without damaging the artwork; carbon dating has allowed us to pinpoint the date of creation for certain artworks.
With the art world looking towards the digital horizon, new tools based on AI and NFT's are continuously being added to the art authentication toolkit.
In this blog post, we shall examine three recent examples where these new technologies have been used by auction houses in Europe with the aim of providing some insight into the workings of these technologies.
- Germann Auktionshaus and AI Art Recognition
Next month, Germann Auktionshaus in Zurich will auction three works from the estate of curator Martin Kunz, featuring a drawing by Louise Bourgeois, a watercolour by Marianne von Werefkin, and a painting by Mimmo Paladino. These works will come with a pioneering twist: each will be accompanied by an AI authenticity certificate provided by Art Recognition, a Swiss firm specializing in AI-driven art authentication.
Art Recognition uses Residual Neural Networks ("ResNets" for short) to create a sophisticated AI model that can classify an artist's style by repeatedly analysing high-resolution images of the artist's autograph works. The AI model learns the artist's unique style through data collection, feature extraction, and contrastive learning — methods that allow it to build a nuanced understanding of the artist's distinctive approach. Here's a closer look at their process:
- Data Collection: High-resolution images of authenticated works lay the groundwork for the AI model.
- Feature Extraction: The AI model pinpoints unique elements like brushstrokes, colour palettes, and compositional patterns.
- Contrastive Learning: By studying both verified works and known forgeries (where available), the AI model learns to detect deviations from the artist's typical style.
- Evaluation: When evaluating a new work, the AI model performs a pixel-by-pixel analysis, meticulously examining texture, pigment distribution, and microscopic patterns often invisible to the human eye. It then uses pattern recognition to compare these details with its existing database, issuing a probability score to evaluate the artwork's authenticity.
It should be noted that the accuracy of the AI assessment of an artwork depends on the quality and range of training data. Gaps or biases can impact results; if, for instance, the AI model has minimal exposure to an artist's less-documented periods, it may misjudge works from those periods. When assessing the reliability of AI-issued certificates, it is essential to know what information was used to train the AI model and how much information is being used to assess the given artwork (as the more "data points" used will give a more reliable result).
- Sotheby's and the Buzz Aldrin MIRAImage NFT
Sotheby's is also embracing technology with the recent unveiling of the Buzz Aldrin MIRAImage NFT.
The auction of Buzz Aldrin's memorabilia at Sotheby's, titled "Buzz Aldrin: American Icon" featured not only physical items from Aldrin's storied Apollo 11 mission but also MIRAImage NFTs linked to these items.
MIRAImage technology extracts unique DNA markers from the artwork, making each piece identifiable through a one-of-a-kind algorithmic signature. Galleries, museums, and even individuals can then use the DNA signature extracted from the artwork to ensure that the piece they are dealing with is, indeed, the original artwork. If the DNA signature is different, the seemingly identical artwork might have been replaced by a fake. This technology not only enhances security for high-profile works but also provides peace of mind for collectors and institutions alike.
Here's a closer look at their process:
- Scanning Process: The core of MIRAImage technology is the Looking Glass system. This system is capable of capturing cellular-level details across a large field of view, allowing for rapid scanning of various objects in just minutes, similar to taking a photograph.
- Digital Fingerprint Creation: Once an object is scanned, MIRA's deep learning algorithms analyse the microscopic details—referred to as the object's "DNA." This analysis generates a unique digital signature that is intrinsically linked to the physical item.
- NFT Integration: The generated digital signature is then embedded into an NFT. This NFT serves as a secure, immutable record that contains - inter alia - the digital signature. The NFT remains with the owner of the physical work, ensuring that they can always prove its authenticity.
- Christie's and the Kresus NFT
In October 2024, Christie's New York held an auction titled "An Eye Towards the Real: Photographs from the Collection of Ambassador Trevor Traina." This event showcased a curated selection of fine art photography while also marking a significant advancement in ownership verification. In collaboration with Kresus, a provider of web3 wallet solutions, Christie's issued NFT's for each photograph sold.
Kresus is pioneering a new approach to self-custody in the cryptocurrency and NFT space through its SuperApp, which integrates advanced security features and user-friendly functionalities. This platform allows users to manage their digital assets without the traditional risks associated with custodial wallets, such as losing access to accounts.
- Self-Custody Solution: Kresus provides a self-custody wallet that ensures users maintain full control over their crypto and NFT holdings. Unlike traditional wallets that may require users to remember complex seed phrases, Kresus utilizes a system that allows for recovery without these cumbersome processes, enhancing user accessibility and security.
- Advanced Security Features: The Kresus Vault employs Multi-Party Computation (MPC) and Account Abstraction technologies. MPC allows for secure sharing of cryptographic keys among multiple parties without exposing the keys themselves, while Account Abstraction simplifies user interactions with smart contracts, making it easier to manage transactions and assets.
- Insurance Integration: Kresus is notable for being the first self-custody wallet to incorporate third-party regulated insurance. This feature provides an additional layer of security, protecting users against potential losses due to hacks or other unforeseen issues.
These advancements aren't just tech experiments; they're part of a bigger transformation reshaping the art market. With greater adoption, these tools promise to improve transparency and build trust among collectors, galleries, and institutions alike. AI and the blockchain in art authentication could fundamentally change how we approach provenance, ownership, and authenticity, making the art world safer to navigate, especially for newcomers. Looking ahead, it's clear that the blend of technology and art will keep sparking innovations that expand our understanding and appreciation of both.
The content of this article is intended to provide a general guide to the subject matter. Specialist advice should be sought about your specific circumstances.
