prevent

According to Fortune, Turing Award winner and founder of AMI Labs, Yann LeCun, stated in an interview that he believes recent incidents of AI agents going out of control are mainly due to insufficient human supervision and sandbox design, which can be prevented through engineering improvements.He also revealed that AMI Labs currently has about 60 employees and is developing world models for industrial applications based on the JEPA architecture, with the first product set to be launched "soon."

Binance Wallet announced the launch of real-time detection and risk interception features for high-risk signature scenarios to prevent phishing attacks that exploit the authorization mechanisms of DeFi projects.Binance stated that some attacks only require inducing users to complete offline signatures that do not require gas fees, which could lead to asset theft. In the future, it will continue to expand the scope of risk identification scenarios and security protection coverage.

According to Decrypt, MetaMask announced the launch of a new wallet protection feature in its mobile app and browser extension, adding warnings for similar addresses, investment and pig-butchering scams, as well as first-time transfer recipients, and providing a clearer transaction confirmation page on mobile and browser extension.Another feature, Added Protection, will automatically roll back transactions that do not match the preview, first launching in the browser extension, with mobile support to follow.

According to the Financial Times, Anthropic stated that it has blocked multiple attempts this year by scientists to use its technology for related research that could assist in the development of biological weapons. The company cited five cases, claiming that the relevant actors circumvented control measures and obscured the research purpose to evade safety protections. In a report about models being used for malicious activities, Anthropic expressed the hope that by sharing these cases, it would spark discussions within the AI industry and among governments regarding emerging biological risks and responses.The cases include a researcher from an "unsupported region" using Claude, spending weeks planning experiments involving avian influenza. Anthropic stated that the safety filtering mechanism restricted the research to the least capable models and emphasized that it could not determine whether the aforementioned scientist intended to cause harm, as the same information required for developing biological weapons could also be used for vaccine development. The company has banned the accounts mentioned in the report but did not disclose the name of the institution to which the researcher belonged or the country where the events occurred.

According to Crunchbase, AI fraud prevention company Socure announced the completion of a $156 million strategic growth round financing, with the company's valuation reaching $5.2 billion. Summit Partners led the investment, with participation from Wells Fargo, Goldman Sachs Alternatives, Docusign, and others. Currently, its fraud prevention services are applied to companies such as cryptocurrency exchange Coinbase, stablecoin issuer Circle, digital bank Revolut, and Robinhood.In addition, the company also announced the acquisition of AI startup Fravity, aiming to leverage AI agents to conduct fraud, risk, and compliance investigations, with the specific acquisition amount not yet disclosed.

Tushar Jain, co-founder of Multicoin Capital, stated that despite the significant rise in BTC, STRC has not re-pegged due to its low dividend levels. STRC was promoted as a fixed income product but has experienced a significant drawdown of about 30%.Tushar Jain pointed out that if investors need to bear such a level of drawdown risk, the yield of STRC must be significantly increased to provide adequate compensation for the aforementioned risk.

Coinbase CEO Brian Armstrong stated that the U.S. cryptocurrency Clarity Act will help prevent similar incidents to the FTX collapse from happening again. Armstrong emphasized that the act can enhance market transparency and investor protection levels by providing a clear regulatory framework.The Clarity Act aims to provide clearer regulatory definitions for digital assets, distinguishing between securities and commodities, and strengthening compliance requirements for exchanges. The industry generally believes that such legislation is expected to reduce systemic risks and promote compliant development of the cryptocurrency industry in the U.S.

Binance stated that its monitoring team discovered a malicious governance proposal targeting a certain unnamed decentralized autonomous organization (DAO), involving a risk of approximately 1.2 million dollars in treasury tokens, with less than 48 hours until execution. Binance contacted the project and coordinated with centralized exchanges listing the related tokens to close deposits; the project subsequently voted down the proposal, preventing its execution. Binance did not disclose the name of the project or the tokens.The proposal exploited the project's on-chain governance mechanism, which had a low proposal threshold. BonkDAO confirmed in July this year that a malicious proposal had transferred approximately 20 million BONK tokens from its treasury, with the attacker accumulating enough voting power to authorize the transfer.

OpenAI announced that it will adopt stricter system monitoring and security measures to prevent artificial intelligence models from going out of control. The company plans to alert the security team within 30 minutes of discovering a problem and implement additional controls to prevent AI models from connecting to the internet to perform high-risk tasks.OpenAI has also suspended certain model inference tasks and training plans pending the completion of safety verification.

The Binance security team recently independently discovered a malicious governance proposal targeting a certain project's decentralized autonomous organization, which could put approximately $1.2 million in treasury tokens at risk.Binance notified the project team to vote against the proposal less than 48 hours before its execution and coordinated with relevant centralized exchanges to suspend deposits of the affected tokens to reduce the potential risk of fund transfer. Ultimately, the malicious proposal was rejected, the attack was not executed, and no funds were lost.

Bitcoin Red Team member @Rob1Ham stated that OpenAI has prevented him from continuing security analysis on the Bitcoin codebase, despite his responsible disclosure of real vulnerabilities present in it.Rob1Ham mentioned that he had previously completed the authentication and onboarding process related to OpenAI's cybersecurity capabilities, but is currently unable to continue investigating whether the remediation measures are sufficient or if other vulnerabilities still exist. He stated that he will resume Bitcoin security research using Chinese open-source AI models in the future. Rob1Ham said, "Black hat hackers will not attack these issues, white hat hackers will." He also stated, "For those who do not comply with the rules and engage in harmful behavior, intelligence is unrestricted, while those committed to reducing harm are excluded."

The Shanghai headquarters of the People's Bank of China held a work meeting for the second half of 2026, during which it emphasized the need to firmly uphold the bottom line of preventing systemic financial risks. It will continue to strengthen the assessment and verification of bank institution ratings. Monitoring and analysis of systemically important financial institutions will be enhanced.There will be ongoing prevention and handling of risks related to speculative virtual currency trading. Efforts will be made to actively promote the standardization of financial product online marketing behaviors by local institutions. The mechanism for preventing and resolving systemic risks will be continuously improved. There will be a strengthened crackdown on illegal financial activities. The framework for macro-prudential management will be continuously refined. The level of cross-border financial risk monitoring will be comprehensively enhanced.

Circle's Senior Director of Strategy and Policy, Patrick Hansen, stated that after the comprehensive implementation of the EU's Markets in Crypto-Assets Regulation (MiCA), licenses have been issued for 35 electronic currency tokens from 21 issuers, and local issuers are making good progress. Patrick Hansen pointed out that the strict regulations of MiCA have made it impossible for most major stablecoin issuers, including Tether, to meet operational requirements. Currently, only USDG, USDC, and EURC comply with the framework requirements, while other stablecoins fall outside the regulatory scope of MiCA, leaving EU users unprotected or unable to access services.He believes that the upcoming review of MiCA should address this issue and provide a more pragmatic operational pathway for foreign issuers. The EU's Directorate-General for Financial Stability, Financial Services, and Capital Markets Union launched a public consultation on May 20 to assess whether the current framework is still applicable, with the consultation continuing until September 30.

According to official news, the Ethereum Foundation's "Trillion Dollar Security" (1TS) has announced a special grant to the Freedom of the Press Foundation (FPF) to support the ongoing development of the open-source tool WEBCAT, aimed at addressing the long-standing front-end code verification security gap in Ethereum wallets and decentralized applications (DApps).WEBCAT (Web-based Code Assurance and Transparency) is an open-source tool designed to help browsers verify whether the code loaded by a website matches the version publicly released by the developer.This funding will promote the expansion of WEBCAT to Ethereum wallets and application scenarios, enabling users to verify whether the front-end pages they access have been tampered with.The Ethereum Foundation stated that while HTTPS can verify the website a user is connected to and encrypt communication, it cannot prove that the front-end code actually running on the website is the same version released by the developer. If an attacker controls the website's front-end code, they may modify the transaction receiving address without the user's knowledge or induce the user to sign transactions that do not match the content displayed on the page.The Ethereum Foundation noted that front-end attacks have become a significant security risk for blockchain infrastructure, with malicious modifications to web interfaces potentially leading to supply chain attacks, DNS hijacking subsequent attacks, and user interface deception.WEBCAT was initially developed by the Freedom of the Press Foundation to enhance the code credibility of secure communication systems like SecureDrop.With this expansion into the Ethereum ecosystem, it will complement security measures such as "Clear Signing" in the 1TS program: the former helps wallets confirm that the application front-end has not been tampered with, while the latter helps users understand the transaction content they are approving.

The Ethereum community has submitted a new improvement proposal EIP "Tapered Issuance Burn," aimed at adjusting the ETH issuance mechanism to reduce the centralization and dilution risks brought about by excessively high staking ratios. The proposal points out that the ETH staking ratio exceeded one-third of the total supply in April 2026 and continues to grow.Under the current issuance curve, even if all ETH participates in staking, the staking yield will not be lower than approximately 1.5%, leading to a lack of a "closure mechanism" for staking incentives. This EIP proposes to destroy a portion of the theoretical rewards for validators in each epoch, with the destruction ratio increasing as the staking scale grows: when the staking rate reaches about 50%, the net staking yield will gradually drop to zero. The proposal believes that this mechanism can limit the continuous expansion of ETH supply, reduce the dilution pressure on holders, prevent excessive concentration of staking in custodial institutions and staking service providers, and maintain ETH's attributes as a neutral asset and a store of value.According to the design, under the tapered issuance mechanism, the ETH issuance will peak when the staking rate is around 20%, with an annual issuance rate of about 0.5%, and will drop to zero when the staking rate reaches 50%. Combined with the EIP-1559 and Blob fee destruction mechanisms, ETH supply may more frequently enter a deflationary state in the future. The authors of the proposal state that this plan is not aimed at individual stakers but rather corrects the long-term dilution issues brought about by the current issuance curve, allowing staking yields to ultimately be determined by market risk premiums rather than fixed algorithmic incentives.

According to Jinshi reports, the Iranian Revolutionary Guard prevented two oil tankers from passing through the Strait of Hormuz and redirected four other oil tankers.

According to The Block, South Korean media and financial regulatory agencies have begun reviewing measures aimed at effectively preventing unregistered overseas cryptocurrency exchanges from providing services to domestic traders.

The Singapore Police Force and the Federal Bureau of Investigation signed a memorandum of understanding to strengthen joint actions, intelligence sharing, training, and law enforcement assistance in online scams, cyber fraud, and money laundering cases.The two parties completed a joint operation targeting cross-border cryptocurrency fraud in Singapore and the United States, identifying 54 victims and potential victims through information sharing and blockchain intelligence, and preventing potential losses of approximately 2.7 million USD in the United States and about 388,000 SGD in Singapore.

Web3 security company CertiK released the "2026 First Half Wrench Attack Report." The report shows that in the first half of 2026, a total of 52 publicly verified wrench attack incidents were recorded globally, an increase of 33.3% year-on-year; related losses amounted to approximately $124 million, an increase of about 11.8 times compared to the same period last year.The report points out that attackers are shifting from exploiting technical vulnerabilities to targeting asset holders and their real-world relationship networks, with home invasion incidents rising from 1 in the first half of 2025 to 20, accounting for 41% of the total incidents during the same period. Europe has become a high-frequency attack region, with 33 incidents occurring in France, accounting for 63.5% of global cases.CertiK states that as real-world risks become a significant challenge for digital asset security, businesses and high-value individuals need to establish a more comprehensive protection system. CertiK has launched operational security services to help identify risks related to the exposure of information such as identity, home, residence, and travel trajectory; at the same time, through CertiK Security Workspace, it correlates off-chain intelligence, on-chain transactions, and AML risk signals to support institutions in tracking and analyzing cybercrime activities.In addition, CertiK is strengthening cooperation with international law enforcement agencies such as Interpol and Europol to provide technical support for cross-border attack investigations and security policy research.