Vitalik Buterin Outlines Potential Crypto + AI Use Cases and Risks
Ethereum co-founder Vitalik Buterin recently explored emerging intersections between crypto and artificial intelligence in an extensive blog post. He argued that certain applications appear to be far more robust than others currently.
Buterin categorized crypto + AI opportunities into four types, ordered by viability:
Most Viable
- AI participates trustlessly in blockchain mechanisms as players.
- Rewarded or penalized by assessed outcomes on-chain.
- Arbitrage, predictions, and identity verification
High Potential
- AI helps users interpret complex Web3 interactions.
- Assist in avoiding mistakes and detecting scams.
- Significant risks remain around adversarial attacks.
Proceed Carefully
- On-chain contracts directly integrate opaque AI components.
- “AI judges” pose numerous integrity and bias risks.
- Crypto techniques may boost transparency over time.
Longer-Term Potential
- Incentivize ethical AI training via crypto protocols.
- Enables query limits and kill switches.
- Ambitiously complex to realize thus far.
Read more: Floki Team Pushes Back On Hong Kong Regulator Warning
The analysis argues that AI acting as automated players within blockchain incentive structures is likely the simplest and most immediately viable model.
Buterin says AI can identify fake accounts
Buterin mentioned that it could involve prediction market information aggregation, identifying fake accounts, assessing data quality, and moreāall tied to transparent on-chain rewards adjudicating performance.
“These applications have promise, both for functionality and for improving AI safety in a way that avoids the centralization risks associated with more mainstream approaches to that problem,” Buterin said.
On the other hand, Buterin mentioned that relying directly on AI components in a smart contract or protocol seems much more dangerous. This is because model integrity can’t be proven, and there are loopholes for attackers.
Read more: DeFi Protocol Abracadabra Finance Suffers $6.4 Million Hack
Buterin thinks that even though using advanced encryption or shared control could make things more reliable later on, users should be careful about using “AI judges” right now.
The Ethereum co-founder suggests that by understanding the good and bad points, his ideas can help users think carefully about how to innovate in a space where two very important industries meet.