Vitalik Buterin, co-founder of Ethereum, has echoed a new caution regarding the potential risks that advancements in artificial intelligence (AI) may pose to the cryptographic methods that secure contemporary blockchains, possibly even before quantum computing becomes a reality.
Buterin’s remarks came in response to a tweet from Ethereum researcher Justin Drake on Wednesday, who urged the blockchain community to enter “bunker mode” as AI advancements could potentially crack the elliptic curve digital signature algorithm (ECDSA) used to protect crypto wallets. Drake suggested that users should begin to gradually transfer their assets to new wallets that do not reveal their public keys.
“While I’m not advising anyone to hastily move their assets today,” Buterin stated, “we must take the threats to cryptography from AI-enhanced mathematics seriously.”
Drake’s apprehensions arose after OpenAI published numerous new mathematical discoveries on a range of topics, including algebra and theoretical computer science, showcasing how rapidly AI is evolving in the field of mathematics.
Recently, the company deployed a team of 10,000 autonomous AI agents to tackle the Navier-Stokes equation—considered one of the most challenging unsolved issues in math and physics—in just 88 hours.
“Recent developments have shown the limitations of human mathematical intuition. Long-standing assumptions have been challenged,” Drake noted, emphasizing that elliptic curves may be particularly susceptible to superintelligent systems.
“Curves possess intricate structures and offer numerous possibilities for clever techniques like Schoof and Frobenius, unlike hashes that are specifically designed to limit algebraic structure,” he explained.
Buterin highlights concerns for quantum-resistant cryptography
Buterin further broadened the discussion to lattice-based cryptography, alerting the community that methods considered resilient against quantum threats could also be compromised by advancements in AI mathematics.
“Up to now, many have perceived the risk as ‘elliptic curves compromised, hashes safe, lattices secure,’” he commented. “However, it’s likely that the concrete security of lattice methods could face significant challenges from the forthcoming breakthroughs in AI mathematics.”
This concern is a key factor behind Ethereum’s focus on a “hash-only” roadmap.
Haseeb Qureshi, managing partner at Dragonfly, echoed the importance of taking these warnings seriously, calling Drake’s statements “a very sobering reminder.”
“The real risk lies not in quantum threats, but rather in conventional mathematics disruptively challenging untested cryptographic assumptions,” he noted on platform X.
Careful transition to new wallets
Alongside these long-term cryptographic considerations, both researchers recommend that cryptocurrency holders mitigate their exposure by keeping funds in addresses that have not disclosed their public keys.
“I suggest initiating a controlled and gradual migration of assets to new addresses,” Drake advised.
He emphasized that larger and more sophisticated crypto holders should be the first to transfer their assets to new wallets, and stressed the need to move any remaining funds to a new address following any transaction.
Related: Nvidia introduces AI safety platform aimed at curbing ‘rogue’ AI agents
Buterin agreed with this precaution, stating that if it is simple to do, utilizing addresses that haven’t been used for transactions is advisable. “However, it is crucial to note that moving funds carries its own risks,” he cautioned.
“Personally, I have faced more losses due to mishandled migrations than from all hacks combined,” Buterin shared.
Drake echoed this sentiment, advising that any migration process should be gradual to avoid unnecessary complications, warning that “a hurried transition could do more harm than good.”
Magazine: Too large to halt: Can a slowdown in AI lead to an economic downturn?
Cointelegraph is dedicated to providing independent and transparent journalism. This article is written in accordance with Cointelegraph’s Editorial Policy and aims to deliver accurate and timely information. Readers are encouraged to verify information independently.
