Building and Scaling dApps with AI and Smart Contracts BBLC Practice Questions (Page 3)

Questions 11–15

1. 11expert · hard

   A decentralized finance (DeFi) dApp uses an AI model to detect fraudulent transactions and automatically freeze accounts. The model runs off-chain and sends instructions to a smart contract. The team is worried about adversarial attacks on the AI model. What is the most effective mitigation?

   • AFreeze accounts only after a human review of the AI's decision.
   • BAllow the smart contract to directly call the AI model's API without any additional checks.
   • CUse multiple independent AI models and require a majority vote before the smart contract executes a freeze.
   • DKeep the AI model's architecture and weights secret to make it harder for attackers to craft adversarial examples.
   Reveal Solution Discussion Select an answer first
2. 12expert · hard

   A healthcare dApp uses an off-chain AI model to predict patient readmission risk. The prediction is sent to a smart contract that triggers insurance payouts. The team is concerned about patient privacy and the integrity of the AI predictions. Which combination of measures best addresses both concerns?

   • AStore all patient data on-chain and use a permissioned blockchain to restrict access to authorized parties only.
   • BUse a trusted third party to run the AI model and manually approve each prediction before it is sent to the smart contract.
   • CSend raw patient data to the AI service and rely on the smart contract's transparency to deter misuse.
   • DEncrypt patient data before sending it to the AI service, and have the smart contract verify a zero-knowledge proof of the prediction without revealing the underlying data.
   Reveal Solution Discussion Select an answer first
4. 13application · medium

   A supply-chain dApp needs to verify product authenticity by analyzing sensor data streams from IoT devices. The analysis requires heavy image recognition on photos taken at each checkpoint, and the results must be permanently recorded on-chain for auditability. The team wants to avoid overloading the Ethereum network with large image files. Which architecture should they implement?

   • ARun image recognition off-chain using an AI service, then submit a hash of the result and a reference to the off-chain data to a smart contract for on-chain verification.
   • BUse a fully off-chain database and only record a timestamp of the analysis, without any on-chain record of the result.
   • CDeploy the AI model as a smart contract on a sidechain and have the main chain call it via cross-chain messaging.
   • DStore the raw images directly on-chain and have the smart contract perform the image recognition logic during transaction execution.
   Reveal Solution Discussion Select an answer first
5. 14application · medium

   A healthcare dApp collects patient data from wearable devices and stores it on IPFS. The team wants to use AI to analyze this data and provide personalized health insights to users. What is the best way to integrate this AI analysis while preserving user privacy?

   • AUse a centralized AI service that processes the data and stores results on-chain.
   • BEncrypt the data on IPFS and give the AI model the decryption key.
   • CRun the AI model on the user's device and only upload encrypted insights to the dApp.
   • DUpload all raw data to the blockchain and let the AI model read it directly from the chain.
   Reveal Solution Discussion Select an answer first
6. 15application · medium

   A logistics startup wants to add an AI-powered route optimization feature to its Ethereum-based dApp. The AI model must analyze real-time traffic data and produce route recommendations, but the company wants to keep the core shipment-tracking logic on-chain for transparency. Which architecture should the team implement?

   • AStore all traffic data on-chain and use an oracle to trigger the AI model every time new data arrives.
   • BRun the AI model off-chain, submit the recommended route hash to the smart contract, and store the full route data on IPFS.
   • CDeploy the AI model directly inside the smart contract using a Solidity library to ensure on-chain transparency.
   • DUse a centralized server to run the AI model and update the smart contract state directly with the final route.
   Reveal Solution Discussion Select an answer first