IC3 researchers published a 155-page survey on June 8 examining how artificial intelligence and crypto can support each other.
The study says meaningful integration remains early and calls for stronger evidence behind claims that blockchain can make AI agents autonomous, identify generated content, or remove model bias.
The paper does not dismiss crypto. It says zero-knowledge proofs, trusted computing, and blockchains can secure AI systems, preserve records, and support machine payments. Yet the researchers argue that these tools address narrower problems than many industry claims suggest.
Crypto wallets automate AI agents without creating autonomy
“AI systems do not become more intelligent by possessing a wallet,” the authors wrote. A wallet can let an agent trade, pay, and access services without approval for each action. Yet people can still change its rules, shut down servers, or block access to supporting systems.
NEW RESEARCH: IC3 just published one of the most comprehensive surveys on AI & crypto.
As AI & agents scale, model-level guardrails aren’t enough. AI needs:
• Trusted data
• Authenticated 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 allow programmable payments. They say blockchain rails may offer neutrality and censorship resistance, but projects must show 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 lets AI systems make swaps and other on-chain transactions under user-defined rules.
Moreover, Robinhood also introduced separate agentic trading and card accounts while keeping agents away from users’ main assets. These controls support IC3’s view that humans remain in charge.
Blockchain records cannot prove who created content
IC3 says blockchains can timestamp a file and preserve a claim about its origin. However, a network cannot inspect an off-chain image, video, or text and decide whether a human or model created it. An outside classifier must supply that judgment.
If the classifier is wrong, the blockchain preserves the wrong claim. Provenance tools may document registered files, but most online content is not cryptographically anchored. The researchers therefore say blockchains protect record integrity, not the truth of the initial statement.
Decentralization does not remove AI model bias
The survey also rejects the claim that decentralized training or governance automatically produces fairer AI. Bias often comes from training data, model design, and inference methods. Moving those processes onto a distributed network does not correct them.
Blockchain can still make selected records visible and broaden participation in governance decisions. Yet the paper says benefits for model quality remain unclear and need real case studies. It also warns that storing large datasets, checkpoints, and inference records on-chain creates cost and scale limits.
Recent product launches show why the debate matters. As previously reported by crypto.news, Solana and Google Cloud launched Pay.sh so AI agents can buy API access with stablecoins per request. IC3 sees promise in such uses but asks builders to prove that crypto offers better cost, access, or resilience than existing payment tools across real-world agent services.