Presented by
Today's News Recap
Defining Decisions - SEC rethinks expanding 'exchange' definition to include crypto, may abandon effort requiring crypto firms to register as exchanges.
Powerful Proposal - US stablecoin bill updated ahead of Senate banking group vote, aims to split power between state and federal authorities.
SOL Stake - VanEck registers spot Solana ETF in Delaware, boosting $SOL enthusiasm; 3.3 million SOL staked.
New here? Ask Copilot up to 2 free queries per day with Messari Basic.
Top Assets
Messari Quarterly Reports
Messari's protocol reports give you a deep dive on the foundation and state of top crypto protocols, including key metrics and notable events. See the complete list of protocol reports here and get a preview of our latest report below.
From Our Sponsor
The power of tomorrow, tokenized.
Uranium.io is the world's first platform for trading uranium, powered by Tezos. ⚡️
Unlike gold or silver, uranium’s market has high barriers to entry, with minimum order sizes of millions of dollars and thousands of lbs.
Tokenized uranium lowers these barriers, making ownership accessible without complex storage. Uranium, the asset powering the AI revolution, is now widely accessible. 🌍
Learn more at https://uranium.io/
Two Bits
AI and Smart Contracts: A Fundamental Mismatch
By: Armita Jalooli
Last time, I covered AI’s reliance on decentralized infrastructure and DePIN’s role in compute, storage, and training. Today, let’s tackle the controversial AI-smart contract relationship - and why they’re fundamentally incompatible despite the hype.
The Core Problem: Deterministic vs. Probabilistic Systems
Smart contracts are deterministic: the same input always produces the same output, ensuring consensus across all blockchain nodes. AI, however, is probabilistic. It learns from data, adapts, and produces variable outputs that change over time.
This key difference creates several issues that break the fundamental principles of smart contracts:
If different nodes receive different AI-generated results, the smart contract cannot execute deterministically, leading to failed transactions or chain splits.
AI models require continuous updates and learning, but smart contracts are immutable - their logic cannot change after deployment.
Without deterministic execution, smart contracts cannot guarantee trustless, verifiable outcomes, undermining their core purpose.
So, embedding AI directly into smart contracts is not feasible - but that doesn't mean they can't work together.
AI's Role in Smart Contracts
Although AI and smart contracts can't fully merge, AI can still enhance smart contract functionality in structured ways:
AI as an Oracle: AI functions offchain and supplies deterministic outputs via an oracle, instead of running within a smart contract. For example, a DeFi lending protocol could use AI to analyze credit risk, but the smart contract itself would only execute predefined rules based on the AI-generated scores.
Offchain AI with Onchain Verification: AI models process complex data offchain, and their outputs are cryptographically verified before use in smart contracts. For instance, a parametric insurance contract could use an AI-powered weather model offchain, but only record a verifiable output (e.g., "Hurricane detected") onchain.
AI for Security & Risk Monitoring: AI can analyze transactions before they reach the smart contract, detecting fraud, wash trading, or exploits. For example, AI models could monitor NFT marketplace transactions and flag suspicious wash trades before execution.
Image credit: The Daily Hodl had the perfect shot.
AI-Powered Smart Contracts: The Unresolved Risks
Even with workarounds, AI-powered smart contracts pose serious risks in ethics, security, and governance:
The Black Box Problem: AI models lack transparency, making it hard to audit decisions. Biased or faulty AI inputs could manipulate smart contracts.
Security Threats: AI is vulnerable to adversarial attacks, where manipulated data leads to false predictions and exploits.
Regulatory Gaps: Who’s liable if an AI-driven contract causes harm - the developer? the AI provider? or the contract owner? Legal frameworks aren’t prepared.
Lack of Standards: Without clear rules, AI-integrated smart contracts remain fragmented, raising concerns about security, fairness, and interoperability.
Final Thought: Due to their fundamental algorithmic differences, AI and smart contracts cannot fully merge. However, AI can still enhance smart contract-based applications when used offchain, as an oracle, or for security analysis. The security, transparency, and legal challenges remain unsolved - and until they are, AI-powered smart contracts will remain one of the most controversial topics in Web3.
Till the next one… 🚀
