Ethereum Foundation says AI agents can find real bugs but triage is the real work

AI hallucinations could lead to scalable botnets, posing significant cybersecurity risks as AI systems gain more autonomous capabilities. The post Researchers warn AI agents could become botnets through hallucinations appeared first on Crypto Briefing.

Lyzr, a startup that builds AI agents for enterprises, used its own AI agent to raise a $100 million round — proof, evidently, that the product actually works.

Widespread API key sharing among AI agents heightens security risks, potentially leading to severe breaches and irreversible damages, especially in crypto. The post 69% of enterprises share API keys among AI agents, creating massive security blind spots appeared first on Crypto Briefing.

The post Ethereum Foundation reveals why AI still fails at finding real bugs appeared on BitcoinEthereumNews.com. The Ethereum Foundation has revealed that the biggest challenge in AI-assisted security research has become proving which reported vulnerabilities are genuine rather than finding potential bugs. Summary Ethereum Foundation says verifying AI bug reports is harder than generating them. AI agents found a real libp2p vulnerability, later disclosed as CVE-2026-34219. The Foundation says human validation and reproducible proof remain essential for protocol security. According to the Ethereum Foundation’s Protocol Security team, recent experiments with coordinated AI agents uncovered real software flaws across systems that Ethereum depends on, but the organization said the majority of the effort now goes into separating valid findings from convincing false positives. The team described the results in a technical post explaining how it has been testing AI agents against systems sof

The Ethereum Foundation has revealed that the biggest challenge in AI-assisted security research has become proving which reported vulnerabilities are genuine rather than finding potential bugs. According to the Ethereum Foundation’s Protocol Security team, recent experiments with coordinated AI agents…

The post AI Agents Could Be Turned Into Botnets Through Hallucinations, Researchers Warn appeared on BitcoinEthereumNews.com. In brief Researchers introduced “Adversarial HalluSquatting,” an attack that exploits AI-generated hallucinations. The technique tricks AI agents into trusting fake repositories or tools that contain malicious instructions. Tests against popular AI coding assistants showed the method could lead to remote code execution in controlled experiments. AI hallucinations may be more than incorrect answers—they could become a way for hackers to compromise computers, according to new research from Tel Aviv University, Technion, and Intuit. In the paper, “Beware of Agentic Botnets: Scalable Untargeted Promptware Attacks via Universal and Transferable Adversarial HalluSquatting,” researchers demonstrated a technique that exploits AI models when they generate fake links to software repositories and other online resources.  “The growing adoption of agentic LLM applications h

Researchers warn AI agents could be tricked into downloading malicious code by exploiting the same hallucinations that cause chatbots to make mistakes.

The post The Next AI Race Won’t Be Won By Smarter Technology. It Will Be Won By Better Judgement. appeared on BitcoinEthereumNews.com. AI agents are rapidly reshaping customer experience. Every recommendation, every checkout and every automated decision is an expression of your brand’s judgement, not just its technology – and that is where the trust equation hits. getty Every major technology shift has forced business leaders to ask a different question. The industrial revolution asked how we could produce more. The internet asked how we could connect more. Mobile asked how we could make every interaction easier. Artificial intelligence is asking something far more uncomfortable. What happens when your organisation no longer scales capability, but scales judgement? That’s the real shift taking place today. While much of the conversation remains focused on faster models, bigger datasets and more capable AI agents, those are rapidly becoming the price of entry. The organisations that cre