Conference Presentation, Fireside Chat, Product Demonstration
Eric Piscini (Hashgraph) & Jonathan Dotan (EQTY Lab): Proof of Trust Rethinking AI ft. DLTBlockchain
- The primary barrier to scaling AI for businesses is identified as the need for trust rather than cost or complexity, with organizations delaying mission-critical delegation until reliable agent verification is established.
- A new evolution in human understanding of trust is anticipated alongside AI development, necessitating a "trust layer" for agents to facilitate value exchange between users and third parties.
- Cryptographic solutions now utilize secure enclaves in CPU and GPU hardware to create immutable, notarized certificates that track agent training data and actions, enabling verification for up to 10 years into the future.
- Prototypes leveraging Blackwell hardware achieve cryptographic proof speeds 400,000 times faster than previous methods, with specific hardware capabilities from Intel and NVIDIA becoming available in 2023.
- Future AI agents will operate dynamically in milliseconds across collaborations rather than at static URLs, requiring lightweight certificates and sovereign guarantees similar to passports to access internet resources.
- Traditional security measures like deep packet inspection or firewalls are deemed ineffective for a future ecosystem with trillions of agents, shifting focus toward decentralized infrastructure to prevent single-entity control.
- AI is positioned as the next "killer app" for blockchain following stablecoins, with plans to register certificates on the Hedera blockchain to support smart contracts for payments, licensing, and access control.
- Technological adoption is expected to follow a trajectory similar to SSL certificates during the e-commerce era, requiring time to achieve full trust and ubiquity despite current breakthroughs.
- Predictions indicate AI agents will soon outnumber humans and drive marketing strategies, with last year marking an inflection point where non-human web activity first exceeded human activity.
- Network activity is forecasted to grow at a compound annual rate of 120%, potentially reaching 2,000 exabytes, driven by an estimated projection of trillions of agents globally.
- The virtual agent market is expected to expand into a $7 trillion physical AI sector, with projections for 37% of all homes to contain humanoid robots by 2050.
- Jonathan anticipates extending the verification technology to mission-critical workflows in the life sciences, public sector, and finance, while Eric notes a new wave of blockchain activity in Europe.
- Marketing efforts are expected to secure guaranteed confirmation of message consumption by AI agents, addressing the scale of agent activity that exceeds human interaction on the web.