New Models of Trust on the Frontiers of Authenticity
summary
Starling Lab spends five years building end-to-end authenticity tools: it ingests five petabytes of genocide video, prototypes a camera-tethered notary that becomes part of the C2PA standard, and ships authenticated Reuters photos (including an archive of Jan 6). Cryptography backs a Rolling Stone investigation using payroll records and early zero-knowledge proofs (which strained GPUs) and a 100-page legal analysis supports an ICC Article 15 submission; a 20 petabytes node for preserved, tamper-evident archives goes live soon.
AI-written summary · about
chapters (9)
- 0:00New trust models in authenticity
- 2:13Cryptography for vulnerable archives
- 5:23Legal credibility with digital evidence
- 8:55Authenticated investigative journalism
- 11:46AI inference challenges
- 15:40Decentralized knowledge networks
- 18:08Risks in open source AI
- 20:38Web 3 for decentralized AI
- 22:0420 petabyte storage node
related talks
30mCapture, Store, Verify with Authenticity by Design
23mWeb 3, Authenticity and the Post-Photographic Age
03mProject Updates
36mDemo & Workshop: Authenticity Tools in Action
10mGenesis Final Round: OpenOrigins
10mGenesis Startup Competition Pitch: OpenOrigins
45mEstablishing Trust in AI: Blockchain
36mSecurity for AI and Multi-Party Collaboration with Confidential Computing (Oh, and...