In an Anthropic simulation test, AI Agents discovered a $4.6 million smart contract vulnerability.

By: theblockbeats.news|2025/12/02 03:15:56

BlockBeats News, December 2nd, Anthropic's latest research shows that AI Agents possess significant on-chain attack capabilities: in a simulation test on truly exploited smart contracts from 2020 to 2025, Claude Opus 4.5, Sonnet 4.5, and GPT-5 together reproduced vulnerabilities worth approximately $4.6 million; while scanning 2849 contracts with no known vulnerabilities, the two models also discovered 2 entirely new zero-day vulnerabilities and successfully simulated profit.

The research points out that the on-chain attack returns of AI have doubled approximately every 1.3 months over the past year, and it has already fully developed the autonomous, profitable vulnerability exploitation capability.

At WEEX, we recognize that the holiday season often brings a different trading experience for many users. As market participation slows, price behavior can feel less predictable and familiar trading rhythms may shift. For traders following AI news today or using AI trading tools, this period often highlights how market structure can influence model performance and short-term signals. Approaching these periods with clear expectations and a disciplined mindset can help traders better navigate seasonal market conditions.

Social Engineering in the Crypto Universe: Safeguarding Your Assets in 2025

Key Takeaways Social engineering, a psychological manipulation tactic, has been the leading cause of crypto asset theft in…

Vitalik Buterin Discusses Grok’s Impact on X’s Truthfulness

Key Takeaways Grok, an AI chatbot, is praised by Vitalik Buterin for enhancing the truthfulness of the social…

Vitalik Buterin Says Grok Keeps Musk’s X More Honest

Key Takeaways Vitalik Buterin believes Grok is a valuable addition to X by challenging users’ preconceptions. Grok’s unpredictability…

Vitalik Buterin Believes Grok Enhances X’s Truth-Friendliness Despite Shortcomings

Key Takeaways Grok, an AI chatbot, is being praised by Vitalik Buterin for making the X platform more…

Palmer Luckey’s Erebor Reaches $4.3B Valuation as Bank Charter Progresses

Key Takeaways: Erebor, a digital bank co-founded by Palmer Luckey, has raised $350 million, bringing its valuation to…

Kalshi First Research Report: When Predicting CPI, Crowd Wisdom Beats Wall Street Analysts

Kalshi’s research shows that the prediction market's judgment of CPI is significantly superior to traditional institutional consensus when unexpected inflation shocks occur

Polymarket Announces In-House L2, Is Polygon's Ace Up?

When top-level applications begin to have the ability to independently support users, traffic, and economic activities, if the underlying network cannot provide additional value, it will inevitably be "backstabbed."

AI Trading Risks in Crypto Markets: Who Takes Responsibility When It Fails?

AI trading is already core market infrastructure, but regulators still treat it as a tool — responsibility always stays with the humans and platforms behind it. The biggest risk in 2025 is not rogue algorithms, but mass-adopted AI strategies that move markets in sync and blur the line between tools and unlicensed advice. The next phase of AI trading is defined by accountability and transparency, not performance — compliance is now a survival requirement, not a constraint.