2024 Toronto Blockchain Futurist Conference

AI Panel:
Moderator: Ron Chan, CO FOUNDER – Inference Labs Inc. 
Speakers:

• Ben-Zion Benkhin, CEO WOMBO 
• Ashley Nandan, ACCOUNT EXECUTIVE, Google Cloud Canada
• Rongui Gu, CO-FOUNDER, CertiK
• Nick Emmons, CO-FOUNDER & CEO, Allora Labs

In an era where artificial intelligence  is not just knocking on our doors but has already made itself at home, the intersection of AI, privacy, and Web3 technologies becomes a pivotal discussion. At the Blockchain 2024 Futurist conference, I attended a panel titled “Artificial Intelligence: Hyperscaling AI, how Web3’s Role is reshaping the world,” industry leaders gathered to dissect these themes. Here’s a dive into the key takeaways:

Decentralization vs. Centralization in AI

Ron Chan, co-founder of Inference Labs Inc., kicked off the discussion by addressing a common query: When will decentralized AI match or surpass centralized systems like ChatGPT? He emphasized that technology itself doesn’t lean towards centralization or decentralization; instead, it’s about trade-offs. Centralized AI might excel in performance in certain areas, but decentralized AI brings sovereignty, diversity, and censorship resistance to AI models, crucial for a democratic tech landscape.

In simple terms, think of centralized AI like a single big library where everyone has to go to read or borrow books (data), controlled by one librarian. Decentralized AI is like having books in many small libraries or even people’s homes, where anyone can lend or borrow, making it harder for any one person to control all the books but also trickier to manage.

The Role of Blockchain and Tokens in AI

Nick Emmons from Allora Labs highlighted how blockchain could revolutionize AI by creating decentralized economic systems where resources like compute power and data can be traded more efficiently.   Imagine a marketplace where not just big corporations, but anyone can trade computing power or data. This setup could make AI development faster and more accessible to everyone, because it smartly uses resources from all participants, much like how a bustling market finds the best deals through buying and selling.

Ashley from Google Cloud Canada highlighted the platform’s strategy of integrating over 150 foundational models, which are pre-trained AI models that serve as a base for building more specialized applications. This approach offers flexibility and fosters innovation by allowing developers to adapt and customize these models for various uses without being confined to a single ecosystem. This move underscores a wider industry trend towards greater interoperability and empowering users with more control over AI tools.

Privacy Concerns and Solutions

Rongui Gu of CertiK highlighted a common problem in AI development: while AI needs lots of data to learn and improve, keeping that data private is crucial. Blockchain technology steps in as a clever fix here, using something called zero-knowledge proofs. Think of zero-knowledge proofs like a magic trick where you can prove you know a secret without actually telling anyone what the secret is. This way, AI can learn from data without ever seeing the sensitive details, keeping everyone’s privacy intact.

Challenges and Constraints

The panel acknowledged several constraints in the AI domain:

• Compute Power: Seen as a new form of currency, the availability and distribution of computational resources remain a bottleneck, although decentralized networks could alleviate this.
• Adoption: Ashley discussed the slow adoption curve of new technologies like AI, comparing it to historical examples like Uber, suggesting that even with available technology, societal and business adoption takes time.
• Security and Open Source: The debate around open-source versus proprietary models in AI touched on security risks. While open-source can accelerate innovation, it also exposes models to potential security threats, making proprietary models more appealing for sensitive applications.

Looking Ahead

What keeps industry leaders awake at night isn’t the lack of interest or potential in AI but rather the direction in which AI governance will evolve. Will AI become a monopolized resource or flourish as a decentralized, public utility? Nick Emmons expressed excitement about steering AI towards a decentralized future, where it serves as a public good rather than a controlled asset.

The integration of AI with Web3 technologies like blockchain is not just a technical merger but a philosophical alignment towards creating a more open, secure, and equitable digital future. As we navigate this landscape, the balance between innovation, privacy, and security will dictate how AI reshapes our world.