Melting Memories — Refik Anadol's EEG-data installation
2018
Artists: Refik Anadol · Year: 2018 · Medium: AI-generated installation — visualizations derived from EEG recordings of memory-recall brain activity · Substrate: EEG hardware recording brain activity during memory-recall sessions; ML-processed data converted into flowing projection imagery · Place: Los Angeles, USA · Institution: Pilevneli Gallery; Parsons School of Design collaboration Refik Anadol, a Turkish-American artist based in Los Angeles, premiered Melting Memories at the Pilevneli Gallery in Istanbul in September 2018. The installation was based on EEG data from patients experiencing Alzheimer's and other forms of memory-loss: Anadol's team collected brain-activity recordings from participants at a Swiss neuroscience lab, ran the EEG data through machine-learning processes that extracted patterns Anadol interpreted as the neural signatures of remembering, and rendered those patterns as large-scale projected imagery of flowing, dissolving, re-forming fields of light. The installation's visual language is painterly: soft colors, slow transitions, a sense of something in the process of being lost. Melting Memories is the piece in which Anadol's signature method — real-time generative projection driven by large datasets processed through neural networks — became fully legible as a specific artistic practice. His subsequent much larger commissions (Machine Hallucinations: NYC, 2019; Unsupervised at MoMA, 2022, elsewhere in this archive) scale up the method across different datasets and different institutional frames, but Melting Memories articulates the grammar at a scale where the argument is still visible. What Anadol argues is that large-scale datasets can be rendered as aesthetic objects with genuine emotional force, that machine-learning processes applied to those datasets produce specifically new kinds of computational imagery, and that the resulting visual experience has a specifically twenty-first-century character that no prior medium could produce. The argument is not unopposed — several critics have objected that Anadol's work aestheticizes data in ways that decontextualize or trivialize the source material — but the scale and commercial success of his practice through the 2020s has made him one of the most visible digital artists of the period. Anadol's studio in Los Angeles currently employs several dozen engineers, designers, and researchers. His work has been commissioned by the Museum of Modern Art, the Dolby Theatre, SFMOMA, the Bilbao Guggenheim, and major public-infrastructure commissions across three continents. He was born in 1985. Lore Anadol's 2018 EEG-data collection was done in collaboration with Dr. Ed Boyden's Synthetic Neurobiology Group at MIT, with neuroscientist Dr. Alex Fornito's team in Melbourne, and with clinical teams at a European Alzheimer's research facility (anonymized in Anadol's public materials for patient privacy). The participants consented to having their brain activity rendered as generative imagery; several subsequently visited the installation and reported, in Anadol's telling, emotional responses that were not reducible to either scientific curiosity or art-historical appreciation. Anadol's studio maintains the raw EEG data in archival form with ongoing consent from the participants; it has been used as source material for several subsequent Anadol Studio works that reference or extend the Melting Memories project. Sources • Anadol, Refik — refikanadol.com (studio archive). • Pilevneli Gallery — Refik Anadol: Melting Memories (exh. cat. 2018). • Anadol, Refik — "Data Paintings" (Leonardo Journal, 2020).
Source: daoracle.com
Part of Digital Art · Watch in the documentary
More from Digital Art
- Ian Cheng — BOB — Bag of Beliefs
- Anne Morgan Spalter — academic practice bridges generative and NFT art
- Robbie Barrat — open-source AI-art pioneer
- Holly Herndon — Spawn — AI vocal model as collaborator (PROTO)
- Sofia Crespo — Neural Zoo and AI-generated biology
- Bitcoin Angel — Trevor Jones's early NFT breakthrough
