Exploring Various Opinions on AI in the Crypto & Blockchain Space

Speaking with builders in the crypto space, we can see that the combination of AI and blockchain technology has two sides: on the one hand, it could make blockchain much more efficient and scalable through AI; on the other, it could make market consolidation, high computing demands, and the need for cost-effective solutions more difficult.

Important insights include using blockchain to improve AI's security and honesty while warning against overhyped apps that don't add value. Let's take a look at some of the differing opinions of different experts in the crypto industry.

The Hype and Reality of AI-Based Tokens

There's no denying the buzz surrounding AI-based tokens, and Tristan Dickinson, Head of Marketing and Communications at dYdX Foundation, believes that the actual benefits of AI are being overlooked due to a greater focus on the popular narrative of AI itself.

Dickinson said "There is a lot of hype around AI-based tokens. What hype can gloss over is the underlying benefits of Artificial Intelligence. When programmed correctly, AI should bring efficiency, scalability, and advancements in blockchain capability."

According to Dickinson, AI, and blockchain could work well together because they both use data and code as their base, creating an excellent environment for AI to grow. Dickinson's words also point to a future in which AI is more commonly used in blockchain, going from being a theory to being used in real projects.

"Blockchain is data—and code-based, creating an ideal platform for AI to thrive. As tangible use cases take over hype cycles, we could see AI weaved into the fabric of most if not all projects; however, caution is needed," Dickinson added, further breaking down his viewpoints.

However, Dickinson's cautionary note about the potential misuse of AI underscores the ethical considerations and the need for responsible deployment of AI technologies. Dickinson said, "AI can be utilized for good just as easily as people can use it for bad." As a result, ethics and good government become even more important in developing and using AI in the blockchain space and beyond.

His balanced view points towards a future where AI and blockchain can work together, as long as careful control is maintained to lower the risks associated with AI's ability to be used for malicious purposes.

Market Consolidation Through Network Incentives

Fraser Edwards, CEO and co-founder at cheqd, illuminates the competitive landscape that AI tokens are navigating. Edwards believes that because AI models need a lot of computer power, flexible and cheap processing tools must be available for AI-based projects to grow.

Edwards broke it down, saying, "For AI tokens in particular, there are forces that will consolidate the market through a fight through network incentives for providers. AI models are compute-heavy and hence require sufficient availability at scale. Furthermore, since they are compute-heavy, there is a heavy focus on optimizing for a low cost."

Since cost may play a factor in the growth of AI-based projects, this could affect how these projects operate according to Edwards. "Providers themselves will naturally head where incentive and real rewards are greatest," Edwards said, continuing to say:

"Hence, there will be rapid competition to collect and retain as much network power as possible through rewards and incentives while simultaneously providing cheap computing to users, meaning protocol revenues will be squeezed. There may be a drive to specialize and niche down as a network, but there will be destruction in the absence of differentiation."

Edwards also draws a parallel to the early days of ride-sharing services like Uber and Lyft, when they had to deal with a "cold start" problem, which is when a startup tries to grow in a market with few participants.

"In short, it will become a classic cold start problem situation where networks or protocols need to amass as much supply-side inventory as possible in anticipation of the demand side. It will be a repeat of Uber and Lyft for AI computing, with all the incentive games being repeated," Edwards said.

Finishing on a positive note, Edward expressed his belief that the market is growing and has a lot of promise, saying, "All that being said, the market is burgeoning with huge scope for growth, so this consolidation may be some ways off."

Blockchain: The Antithesis and Savior of AI

Scott Dykstra, CTO & Co-founder of Space and Time, a web3 data warehouse, offers a nuanced perspective on the relationship between AI and blockchain. He describes blockchain as a counterpoint and a crucial ally to AI, especially in the emerging generative era dominated by deepfakes and fraudulent content.

Dykstra explained, "Blockchain is both the antithesis to AI and, conversely, the thing that it needs the most. We all know we're entering the new generative era of the internet, where fraudulent content and deepfakes can be created in seconds."

"Two of the most obvious use cases for AI x blockchain are cryptographic watermarking for on-chain audit trails of content provenance and using ZK to verify the integrity of a model's training data."

Dykstra continued to break it down, saying "But it's important to make a distinction between real use cases where blockchain technology is used to secure and verify AI, and protocols that are building an AI framework simply because they see the writing on the wall of where the industry is headed.

However, he also distinguishes between genuine use cases and the trend of building decentralized AI frameworks for their own sake. The real value lies in delivering cost-effective and rapid AI inference to end users, a goal currently hindered by the increased costs and complexity of verifying AI models on the blockchain.

"Building decentralized AI for the sake of decentralized AI doesn't necessarily provide value to the end user. At the end of the day, the end user wants the cheapest and fastest AI inference possible," Dykstra said.

Dykstra also cites cryptographic watermarking and zero-knowledge (ZK) proofs as two ways to ensure the integrity of training data as an example of how AI and blockchain might work together.

Dykstra said, "The challenge is that verifying AI models in production, whether using optimistic, consensus, or ZK verification methods, increases the cost and complexity of running AI inference. This piece has to be solved before we can move forward and find real adoption of blockchain as a tool for AI,"

Conclusion

When AI and blockchain are used together, it can lead to new ideas and protocols, but it can also cause some challenges, i.e., high costs and high computing power, as mentioned by Edwards.

Suppose projects focus on cost-effectiveness and scalability when installing AI models. In that case, they must balance offering incentives for network involvement and ensuring that computer resources are accessible cheaply. The ideas shared by Dickinson, Edwards, and Dykstra shed light on some of the challenges and opportunities ahead.