Ethereum Researcher Warns of AI Threat to Crypto Security

In a warning to the crypto industry, Justin Drake of the Ethereum Foundation has put forward the case that artificial intelligence poses a threat to security, with private key exposure and ensuing devaluation of digital assets chief among them. The researcher’s position has been the subject of some debate in blockchain circles.

AI Advances Pose New Risks to Cryptography

Drake says that advances in neural networks and mathematics will eventually give hackers the means to extract private keys from public addresses in a matter of months, undermining what is supposed to be the bedrock of blockchain security. He emphasized that a computational cluster with enough power could potentially make it possible to access keys using old elliptic curve cryptography in less than a week.

The upshot of these vulnerabilities is that a successful hack would put huge quantities of cryptocurrency at risk and cause the value of Bitcoin and similar assets to fall rapidly. For Drake, it is imperative that the industry prepares for AI’s impact on crypto security. To that end, he has advised a proactive migration to quantum-resistant hash-based cryptography and moving funds to new addresses altogether.

Expert Advice: Migration and Upgrades Needed

Vitalik Buterin, who co-founded Ethereum, agrees with Drake but has cautioned the community not to panic and trigger mass migrations. A hasty transfer, Buterin observes, could do more financial harm through human error than the vulnerability itself. He pointed out that it is important to keep a close eye on lattice-based cryptography and other methods that are theoretically vulnerable to attack.

Potential Impact on Crypto Markets

If attackers manage to extract private keys from elliptic curve cryptography, the fallout for the market could be considerable. A breakthrough of that kind would be enough to cause a sharp decline in the price of Bitcoin and other assets, Drake cautioned.

As the Incrypted channel has noted, there is no denying the reality of the AI threat, but what is needed is a response that is as measured as it is coordinated, in order to mitigate risk and avoid the kind of damage that user error can cause.