The Problem With AI and How Web3 Can Solve It - FOGcast#49

In this episode of the “Future of Gaming Podcast” by FOGDAO, host Nico Vereecke is joined by Devin Becker, the gaming lead at NAMI, Tim Cotten, the co-founder and CEO at Scripted, and David Bennahum, co-founder and CEO at Ready Games. The discussion focuses on the intersection of AI, Web3 technologies, and their role in the future of the internet.
Key Discussion Points:
1. Dominance of "The Big Three" in AI: David raises concern over the concentration of AI power and data among Facebook, Google, and Microsoft/OpenAI, and explores the ethical implications of such centralization, particularly in terms of data ownership.
2. Decentralizing AI with Web 3.0: The discussion investigates how Web 3.0 and blockchain technologies can potentially decentralize AI, disrupting the control of the "big three" and enabling a distributed ownership model.
3. AI and Autonomous Agents: David explains the concept of autonomous agents in AI and how smart contracts and Web 3.0 could give these agents the legal autonomy to operate effectively. The importance of decentralization in AI models to prevent monopolies is emphasized.
4. Potential Risks of Combining Autonomous Agents and Blockchain: Devin Becker identifies the irreversibility of blockchain transactions and the danger of bad actors exploiting this system as major concerns. He also highlights the risk of identity theft in decentralized systems.
5. Reputation Management with Blockchain in AI: Tim Cotten speaks on the potential benefits of blockchain for reputation management and verification of participation and value in distributed AI training models. He warns that the full potential of blockchain and AI convergence can only be realized when the technology is more user-friendly.
6. Decentralization of AI Models: The conversation delves into the technical feasibility of decentralizing AI models in a world where autonomous agents are part of our everyday lives. David suggests viewing AI as a common utility rather than proprietary technology of a handful of companies. 
7. Subsidized AI and Fragmentation of Technology: The discussion touches upon the concept of subsidizing AI costs through in-kind donations, the importance of making technology accessible and fragmented, and the idea of an open-source entity that can be fragmented further to prevent centralization.
8. Attribution and Accountability in AI Models: The group discusses the need for a system to attribute contributions to AI models, and a verifiable history of data evolution. The role of blockchain in these areas is explored, as well as the importance of AI accountability in a signed environment.
9. Handling New AI Requirements: Concerns are raised about the involvement of third-party services and the need for a standard in signing AI models to manage reputational and security risks.
10. Deep Fakes and Identity Protection: The conversation turns to the topic of deep fakes and the potential repercussions of their widespread use. The necessity of protecting identity to prevent deep fake creation and the potential for a centralized ID system are discussed.
Enjoy!Join the FOGDAO DiscordWatch the episode on YouTubeFollow the Future of Gaming: Twitter | LinkedIn