IC3 researchers published a 155-page study on June 8 exploring how artificial intelligence and crypto can support each other.
The study say meaningful integration remains early and calls for stronger evidence behind claims that blockchain can make AI agents autonomous, identify generated content, or remove model biases.
The newspaper does not reject crypto. It is said that zero-knowledge proofs, trusted computing and blockchains can secure AI systems, preserve data and support machine payments. Still, the researchers argue that these tools address broader issues than many industry claims suggest.
Crypto wallets automate AI agents without creating autonomy
“AI systems do not become more intelligent by owning a wallet,” the authors wrote. A wallet allows an agent to trade, pay, and access services without requiring approval for every action. Yet people can still change the rules, shut down servers or block access to support systems.
NEW RESEARCH: IC3 just published one of the most comprehensive studies on AI and crypto.
As AI and agents scale, model-level guardrails are not enough. AI needs:
• Reliable data
• Verified workflows
• Verified execution
• System level securityChainlink solves this. https://t.co/noOBWHMRRl
— Chainlink (@chainlink) June 8, 2026
The researchers also note that centralized financial systems can enable programmable payments. They say blockchain rails can provide neutrality and resistance to censorship, but projects must deliver measurable benefits over centralized alternatives.
“Automation should not be confused with autonomy,” the paper said.
As previously reported by crypto.news, MetaMask launched its Early Access Agent Wallet on June 8. It allows AI systems to perform swaps and other on-chain transactions according to user-defined rules.
Additionally, Robinhood also introduced separate trading and card accounts for agents, while keeping agents away from users’ most important assets. These controls support IC3’s view that people remain in control.
Blockchain records cannot prove who created the content
IC3 says blockchains can time stamp a file and maintain a claim about its origin. However, a network cannot inspect an image, video, or text off-chain and decide whether a human or a model created it. An external classifier must make that judgment.
If the classifier is wrong, the blockchain retains the wrong claim. Provenance tools can document recorded files, but most online content is not cryptographically anchored. The researchers therefore say that blockchains protect the integrity of documents, and not the truth of the original statement.
Decentralization does not remove the bias in the AI model
The research also rejects the claim that decentralized training or governance automatically produces fairer AI. Biases often arise from training data, model design, and inference methods. Moving these processes to a distributed network does not correct them.
Blockchain can still make select records visible and broaden participation in governance decisions. Still, the article says the benefits to model quality remain unclear and need real case studies. It also warns that storing large data sets, checkpoints, and inference records on-chain comes with cost and scale limitations.
Recent product launches show why the debate matters. As previously reported by crypto.news, Solana and Google have launched Cloud Pay.sh to allow AI agents to purchase API access with stablecoins on a per-request basis. IC3 sees promise in such uses, but asks builders to prove that crypto offers better cost, access, or resiliency than existing payment instruments in real-world agent services.