Should AI Be Decentralized, And if So - How?
Combined Session
Wednesday, June 05, 2024 14:30—15:30
Location: B 05
Log in to download presentations
Wednesday, June 05, 2024 14:30—15:30
Location: B 05
Watch the video
We stand at the brink of an AI revolution, teeming with limitless prospects for an AI-enhanced prosperous future. These possibilities stir a mix of excitement and apprehension. The pivotal question is: how do we effectively harness AI's immense power? This query resonates with a diverse audience, ranging from Hollywood scriptwriters to European political figures. In this presentation, distinguished technologist Wenjing Chu will explore a Socio-Technical approach, which integrates social and technical system designs. His research points to a decentralized approach to optimize outcomes of AI development and adoption by bolstering human agency and leveraging both market dynamics and policy frameworks. Focusing on the decentralization of AI, Wenjing will also delve into the groundbreaking work that he is leading on the Trust Spanning Protocol (TSP) specification and the Authentic AI report. He will discuss their significant roles in redefining digital identity and wallets for the AI era.
Watch the video
Decentralized AI can enhance transparency and accountability by distributing data, computation, validation, optimization, and execution across multiple nodes, thus preventing a concentration of power that could threaten individual freedoms.
It can also foster innovation by allowing third-party developers to verify the data and algorithms that an AI system uses.
Technologies such as blockchain and federated learning are at the forefront of this revolution. Blockchain's transparency, immutability, and decentralization can enhance trust in data-driven decision-making.
Federated learning, on the other hand, allows for the training of machine learning models on decentralized data, thus preserving privacy and reducing the need for data centralization. In this panel session we will discuss the technical foundations and the pros and cons of decentralized AI.
