Ethereum Foundation Wants To Use AI to De-Risk ETH Ecosystem

The move aims to address rollup hacks.

By: Pedro Solimano Loading...

 image of robots fixing a codebase, designed in a minimalistic style with neon colors

The Ethereum Foundation wants to stave off hacks by using Artificial Intelligence (AI), according to one of their main researchers.

“Rollup hacks in the tens of billions [of dollars] is a top systemic risk for Ethereum,” said Justin Drake, one of Ethereum’s leading developers, referring to the layer-2 scaling solutions that bundle transactions and help reduce network congestion and high fees. “We want to use AI to identify catastrophic vulnerabilities that could lead to such zk-rollup hacks.”

AI can help Ethereum researchers find bugs and potential exploits at a time when the codebase of Ethereum-based chains becomes increasingly complex as new technologies, including zero-knowledge proofs, are implemented.

Zero-knowledge (ZK) rollups are scaling solutions that use the minimum data needed to verify transactions on blockchains like Ethereum.

Drake told The Defiant “we hope to derisk zkEVM implementations as zk-rollup progressively remove training wheels,” adding that AI could be a fantastic tool to accelerate automatic bug finding of SNARK circuits, referring to “Succinct Non-Interactive Argument of Knowledge”–a term for cryptographic proof that allows one party to prove it possesses certain information without revealing that information

While hacks in crypto are declining, they continue to drain hundreds of millions of dollars from the blockchain ecosystem. Hackers stole $336 million in crypto in the first quarter this year, a 23% drop from the same time last year, according to the latest report by Web3 security firm ImmuneFi.

The firm also reported that Ethereum surpassed BNB Chain as the most targeted network by attackers, and that 100% of exploits took place in DeFi. There is currently $55 billion of assets held in DeFi on Ethereum, according to DeFiLlama, an attractive target for hackers.

With these numbers, along with the increasing number of users coming on chain, adopting additional tools like AI to protect users can’t come soon enough.

Crypto Losses to Hacks in Q1 2024 - Immunefi

Crypto Meets AI

There's actual substance behind the mashup between crypto and AI, experts say. Tokens at the intersection of the two have ballooned from $9.2 billion in Dec. 2023 to $20 billion today, led by Render and Bittensor, just two of a plethora of AI-crypto assets in the top 100 by market capitalization on CoinGecko.

Render is up 225% in the past year, and trades for $7.82, with a market cap just over $3 billion. Bittensor’s TAO token has enjoyed even more success, surging 800% in the past year, trading for $408 and a $2.7 billion market cap.

Extra Set of Eyes

Kenneth Francis, co-founder of Noometic AI–a company that marries machine learning with the creator economy–is bullish on the Ethereum Foundation’s use of AI.

He explained to The Defiant that since people don’t have the attention spans or the necessary comprehension of code bases, “it's of big value” to add AI as what he calls an “extra set of eyes.”

Drake agreed, explaining that the Ethereum Foundation is ready to have people dedicated to AI.

In fact, on April 11, the developer posted on X, “love AI and Ethereum? EF research is hiring,” adding an email address for interested developers to apply.

“Zooming out, we're hoping to stay ahead of the curve with what may become the most powerful technology of our times,” he told The Defiant. “The EF has 300 people—time to have someone dedicated to AI opportunities.”