AI summary by TwinMind
How AI Art Could Enhance Humanity’s Collective Memory | Refik Anadol | TED
TED
5m 5s
Project Introduction - Coral Dreams
- Artificial corals displayed, not real - created by generative AI algorithm
- Algorithm trained with 100+ million coral images
- Global coral reef death due to climate change making seawater less habitable
- Future may only have coral simulations in virtual worlds
- Project aims: use AI to create artificial realities while preserving disappearing nature
Artist Background & Methodology
- Media artist and director with 7-year generative AI collaboration experience
- Team trains machine-learning algorithms using large, focused, publicly available datasets
- Visualizes "humanity's collective memories" - nature, urban, culture
- Post-pandemic focus: compile largest datasets to artificially preserve nature
- Optimistic about generative AI's potential for memory enhancement
Technical Capabilities & Examples
- Artists can utilize AI to represent nature and remember immersion feelings in digital age
- Generative AI creates training possibilities with: images, sound, text, scent data
- Floral Dreams example:
- Algorithm trained with 75+ million floral images
- Covers 16,000 species
- Used 500,000+ scent molecules to create dream scents
Vision for Living Archives
- Concept: walkable living archive universe
- Constantly reimagined, reconstructing forms, patterns, colors, scents
- Life increasingly rooted in digital worlds
- Blurring boundaries: physical/virtual, technology/nature
- Generative AI creates new realities, projects onto reality through possibility space
Future Innovations & Challenges
- Large language models = beginning of long innovation journey
- Future exploration: hyper models, text-to-image-to-sound capabilities
- Key challenge: providing models with original data
Glacier Dreams Project
- Decision: avoid existing models, collect original data
- Data collection types: image, sound, scent, climate data
- Iceland expedition: first destination to capture glacier narratives
Indigenous Collaboration - Yawanawa Tribe Project
- AI capacity mapping onto complex human wisdom/consciousness in nature
- Question: using AI to preserve/learn ancient nature knowledge
- Collaboration details:
- Location: Brazil, Acre, Amazonia rainforest
- Partners: Chief Nixiwaka, Putanny (cultural preservation/ecological sanctuary leaders)
- Project: respectful co-creation, open-source AI rainforest model
- Capability: reconstruct extinct flora/fauna based on tribe's collective knowledge
- Goal: bring ancient wisdom to society respectfully
Ultimate Vision
- AI as mirror reflecting collective memories of all humanity
- Use AI to connect people across ages and cultures
