Cultural Heritage and AI
summary
Builds a small pipeline that actually works: gathers images into AWS S3, trains a classifier with Amazon Rekognition on about 130 images and correctly identifies regional sari styles such as bandhani, paithani and banarasi. Adds OCR and computer vision for collections, connects an LLM to curated knowledge bases to improve generative prompts, then uses PartyRock to prototype designs and commissions a real sari for $300 AI sari. Raises concrete risks: dataset scraping, attribution loss, and artisan livelihoods.
AI-written summary · about
chapters (10)
- 0:00Introduction and background
- 2:37From appropriation to conservation
- 5:51Textiles and cultural symbolism
- 10:24Digitizing cultural collections
- 13:36AI app for sari identification
- 16:44AI in cultural heritage ups and downs
- 18:53Artisan textiles and mechanization
- 19:33Generative AI for sari designs
- 21:22Meaningful textile symbolism
- 24:10Conclusion and call to action
related talks
25mPersonalizing Your Images with AI-Powered Features
24mAddressing Tomorrow's Problems with Today's AI
27mBuilding Generative AI features
24mKeynote: Building India’s Sovereign AI Stack w/ Local Talent & Open Global Collab
26mInfinite AI
20mExplore what’s possible with new AWS generative AI services
15mAI as a Service: The Next thing for Big Tech
34mEthical framework for AI : Creating AI powered products that empower people