Ethereum Uses AI to Find Bugs and Strengthen Security
The same capabilities that expose software weaknesses can strengthen Ethereum security through bug detection and formal verification.
Every Zipp story tagged #Cybersecurity, newest first.
The same capabilities that expose software weaknesses can strengthen Ethereum security through bug detection and formal verification.
The pledge covers cyber, biological and chemical risks, but leaves audit selection, disclosure and implementation timelines to the companies themselves.
The case puts alleged proceeds from a major crypto hack alongside a striking claim about how some of the funds were spent.
The response adds approval layers and resets credentials, but the affected third-party security functions will remain disabled until remediation is complete.
The scale of the investigations signals a security burden extending well beyond the incidents publicly known, raising the stakes for AI oversight and deployment.
The former Australian prime minister’s criticism puts leadership and security oversight at the center of scrutiny facing OpenAI.
Bitget says its $464M User Protection Fund covers the reported loss, but withdrawals remain frozen as investigators examine the attack and an unconfirmed North Korea link.
The attribution remains unconfirmed, but the alleged DPRK connection raises the stakes for exchange security and scrutiny around the $352M loss.
The reported loss puts exchange custody controls and market confidence back in focus.
The incident puts autonomous AI oversight under pressure as Australia investigates access beyond intended permissions and reviews new duties for developers.
The reported attack chain reaches from Safari to kernel-level access, putting private keys and seed phrases stored on affected devices at direct risk.
The episode shows how coding agents can compress exploit timelines while expanding the reach of compromised accounts through connected developer tools.
The episode exposes how leaked credentials and unintended internet access can turn an AI security exercise into a real-world breach risk.
The breach exposed read-only exchange APIs and trading data, underscoring how infrastructure and operational weaknesses are driving a disproportionate share of crypto losses.
The campaign spans more than 100 countries and regions, putting crypto users at risk through a malware operation linked to North Korea.
Public ledgers are becoming durable malware coordination layers, forcing defenders to monitor wallets, smart contracts and RPC traffic instead of relying only on server takedowns.
The episode highlights the dual-use risk of frontier AI and raises pressure on labs to harden models against cyber abuse.
The episode underscores the dual-use risk of frontier AI: capabilities built for software work can also help researchers probe corporate defenses.
The Ethereum co-founder argues AI that can prove theorems can also formally verify software security, and that staying in crypto is itself a bet on defenders winning the AI arms race.
The incident makes containment a policy and deployment issue, not merely a technical feature of model testing.