🤖 AI × Crypto🔴 BearishActionable

OpenZeppelin finds data contamination in OpenAI’s EVMbench

CoinTelegraph – AI|Cointelegraph by Brian Quarmby|March 3, 2026 at 04:18 AM|2 views

Image via CoinTelegraph – AI

🤖AI Summary

OpenZeppelin discovered significant flaws in OpenAI's EVMbench dataset, including data contamination from training leaks and at least four incorrectly classified high-severity vulnerabilities. This finding raises concerns about the reliability of AI tools used for blockchain security auditing.

Key Takeaways

→OpenZeppelin identified training data leaks contaminating OpenAI's EVMbench dataset.
→At least four high-severity vulnerability classifications in EVMbench were found to be invalid.
→The contamination raises questions about AI reliability in blockchain security applications.
→This discovery highlights potential risks of using AI tools for smart contract auditing without proper validation.
→The findings suggest need for more rigorous testing of AI benchmarks in cryptocurrency security.

#openzeppelin #openai #evmbench #data-contamination #ai-security #blockchain-audit #vulnerability-detection #smart-contracts #ai-reliability #security-tools

Read Original →via CoinTelegraph – AI

Act on this with AI

Stay ahead of the market.

Connect your wallet to an AI agent. It reads balances, proposes swaps and bridges across 15 chains — you keep full control of your keys.

Connect Wallet to AI →How it works

AI × Crypto16h ago

World War III Scenario: Which Crypto Would Suffer the Most? (4 AIs Respond)

Four AI models analyzed a hypothetical World War III scenario to identify which cryptocurrencies would be most vulnerable to massive price declines. The analysis suggests certain tokens could potentially plummet by 90% in such extreme geopolitical conditions.

AI × Crypto1h ago

TraderBench: How Robust Are AI Agents in Adversarial Capital Markets?

TraderBench introduces a new benchmark for evaluating AI agents in financial markets, combining expert-verified static tasks with adversarial trading simulations. The study found that 8 of 13 tested AI models showed minimal variation across market conditions, indicating they rely on fixed strategies rather than adaptive market behavior.

AI × Crypto8h ago

Core Scientific revenue drops in Q4 as bitcoin miner pushes into supplying data center infrastructure

Core Scientific reported declining Q4 revenue from bitcoin mining operations while significantly expanding its data center colocation business. The company's colocation revenue surged to $31.3 million from $8.5 million in 2024 as it diversifies away from pure crypto mining toward infrastructure services.