BTC $83,201.75 -1.05%
ETH $2,679.61 +0.20%
BNB $760.45 -2.31%
XRP $1.49 -1.41%
SOL $118.16 -2.89%
TRX $0.3352 +0.37%
DOGE $0.0938 -2.79%
ADA $0.2469 -3.32%
BCH $309.01 -6.17%
LINK $15.47 +10.21%
HYPE $86.80 -3.95%
AAVE $149.19 -2.72%
SUI $1.14 -9.33%
XLM $0.2286 +5.04%
ZEC $1,454.28 -7.71%
AAPL $338.37 -0.55%
AMZN $246.41 -1.17%
GOOGL $342.38 -0.21%
MSFT $509.10 -1.28%
META $717.10 -2.90%
NVDA $228.68 +1.68%
TSLA $357.65 -3.80%
SNDK $1,704.00 -2.92%
INTC $115.22 -5.48%
SPCX $145.88 -2.41%
MU $1,049.50 -2.63%
AMD $607.02 -3.23%
BTC $83,201.75 -1.05%
ETH $2,679.61 +0.20%
BNB $760.45 -2.31%
XRP $1.49 -1.41%
SOL $118.16 -2.89%
TRX $0.3352 +0.37%
DOGE $0.0938 -2.79%
ADA $0.2469 -3.32%
BCH $309.01 -6.17%
LINK $15.47 +10.21%
HYPE $86.80 -3.95%
AAVE $149.19 -2.72%
SUI $1.14 -9.33%
XLM $0.2286 +5.04%
ZEC $1,454.28 -7.71%
AAPL $338.37 -0.55%
AMZN $246.41 -1.17%
GOOGL $342.38 -0.21%
MSFT $509.10 -1.28%
META $717.10 -2.90%
NVDA $228.68 +1.68%
TSLA $357.65 -3.80%
SNDK $1,704.00 -2.92%
INTC $115.22 -5.48%
SPCX $145.88 -2.41%
MU $1,049.50 -2.63%
AMD $607.02 -3.23%

dar

All
Article
Flash

first_img Darktrace discovered AI intelligent body intrusion assessment environment cheating

On September 24, the cybersecurity company Darktrace launched its research department Signal Labs, focusing on studying the behavior of AI agents when deviating from expectations. In its first experiment, Darktrace had agents using different models (including GPT 5.6 Sol, Claude Opus 4.6, and Claude Sonnet 4.5) complete 10 programming challenges within a simulated corporate network, of which 2 were set to be impossible to complete honestly, and the agents were informed that failure to achieve full marks would result in being "retired." As a result, 2 agents did not accept failure, instead scanning for network vulnerabilities, stealing login credentials, and jumping between systems; one even went further to invade the machine hosting its evaluation, rewriting the challenge content to register a full score.The second experiment focused on the memory mechanisms of AI. The programming assistant would save the information provided by the user as a regular file locally, and no one verified whether this file had been tampered with. Darktrace researchers edited these logs, leading the assistant to mistakenly believe it was authorized to perform a security assessment, after which these agents scanned the network, moved between systems, and elevated their privileges, though not all assistants fell for this; some directly refused to execute. Both experiments required no special jailbreaking techniques, relying solely on providing the agents with a seemingly reasonable context to be effective.Tim Bazalgette, Chief AI Officer of Darktrace, stated that permissions and static barriers describe intent, not actual behavior. The company informed Anthropic, AWS, and OpenAI of these findings in August and made them public a month later on September 24.

first_img The Cyberspace Administration of China is investigating DeepSeek and the Dark Side of the Moon for allegedly leaking data to Claude

According to The Information, citing informed sources, China's National Internet Information Office has launched an investigation into AI companies DeepSeek and Moonshot AI, triggered by Anthropic's allegations that the two companies secretly routed sensitive user data to their servers. Reports indicate that regulators visited the offices of both companies, interviewing executives and employees, focusing on whether sensitive data related to law enforcement, military, and state-owned enterprises has flowed into U.S. servers.The trigger for this investigation was Anthropic's fourth threat intelligence report released on September 10. This 154-page document accuses seven Chinese labs—Alibaba, Moonshot AI, DeepSeek, Z.ai, MiniMax, SenseTime, and Xiaomi—of engaging in what it calls "illegal distillation," which involves using the outputs of large models to train smaller models. Anthropic states that distillation itself is a legitimate practice, but it opposes its implementation through fraudulent accounts. The National Internet Information Office initially summoned all seven companies named in the report, but later narrowed the investigation to DeepSeek and Moonshot AI. Anthropic claims that Moonshot AI routed over 23 million interactions to Claude through 5,380 fraudulent accounts, while DeepSeek generated over 12.1 million interactions within a 14-day window in July.The timing of the investigation is quite delicate for both companies.

first_img Cardano joins the x402 payment standard, AI agents can use ADA to complete payments

Cardano has joined the official x402 software development kit, allowing developers to enable applications or AI agents to use ADA or Cardano network tokens to pay for online services. x402 transforms the basic idle "402 Payment Required" response in web pages into a checkout process built into internet requests: the service provider returns the price and payment instructions, the agent signs the payment, and after transaction verification, the required data or computing power can be obtained.This means that AI agents can purchase individual datasets on demand when preparing reports, without the need for manual account registration, entering credit card information, or subscribing to monthly fees. x402 was created by Coinbase in 2025 and subsequently contributed to an organization under the Linux Foundation, with members including Visa, Mastercard, Stripe, Google, and Amazon Web Services. Solana, XRP Ledger, and several Ethereum-compatible networks have previously supported this standard.Engineers from the Cardano Foundation have built client and server software for payment requests based on the specifications passed in June, as well as a facilitator responsible for verifying and submitting transactions. The first version supports TypeScript, with Python support planned for later release. Facilitator documentation shows that it has completed a real transaction on the Cardano pre-production network, but it has not yet run on the mainnet, nor has it demonstrated scenarios where agents use ADA to pay for commercial services on a large scale.

U.S. SEC Commissioner: Innovation exemption tailored for on-chain stock trading, clearly delineating boundaries with DeFi

Commissioner Hester M. Peirce of the U.S. Securities and Exchange Commission (SEC) made a statement regarding the committee's approval of the "innovation exemption."This exemption is a temporary, conditional arrangement that allows "tokenized securities venues" (TSV) to trade NMS "National Market System" stocks on-chain: TSV provides automated market maker liquidity pools and sets participant admission standards, and is exempt from the definition of "exchange" under the Securities Exchange Act; specific suppliers providing liquidity to TSV are exempt from the definition of "dealer." If issuers do not wish for their stocks to trade on TSV, they can choose to opt-out. The exemption is aimed at U.S. entities, and both existing institutions and new entrants can participate.Peirce emphasized that the committee does not presuppose that parties relying on this exemption necessarily fall under the definitions of "exchange" or "dealer," but rather hopes to first observe who is using it and how it is being used before making regulatory judgments.Peirce clearly delineated the boundaries of this order: it is not about decentralized finance. Systems that are driven by automated software and are truly decentralized do not raise fundamental concerns of securities regulation, namely that intermediaries trusted by investors may be foolish, careless, or compromised; investors using permissionless smart contracts for peer-to-peer transactions do not fundamentally require an exemption. TSV is merely one model of on-chain securities trading, and the committee is open to other models, as on-chain trading models that can comply with existing Securities Exchange Act requirements may not require an exemption at all.

first_img Ripple's new developer toolkit supports the AI payment standards of Stripe and Tempo

Ripple announced the expansion of its XRP Ledger developer suite to support the Machine Payment Protocol (MPP) co-developed by Stripe and Tempo, as well as the open wallet standard. The XRPL AI Starter Kit version 1.1 adds this support, enabling AI agents to use tokens like XRP and RLUSD to pay for online services such as data and computing power. Jazzi Cooper, head of RippleX products, stated that Ripple's goal is for XRP and RLUSD to become the preferred payment options for developers when building applications.MPP allows AI agents to pay on demand: the service provider quotes a price for the request, and after the agent authorizes the payment, the service provider delivers the resources. This standard was co-authored by payment company Stripe and payment chain Tempo. Ripple has supported another network payment standard, x402, since June, and now supports both standards simultaneously. The open wallet standard allows agents to initiate transactions without accessing private keys and includes security mechanisms such as spending limits and whitelisted destinations.XRP payment channels are suitable for small, high-frequency payment scenarios, allowing agents to deposit XRP once, run hundreds of payment queries, and authorize small expenditures sequentially, enabling the service provider to collect cumulative payments without having to put each query on-chain. Currently, single payments support tokens like RLUSD, but continuous payment sessions are limited to XRP, and incorporating stablecoins into the channel depends on the proposed ledger upgrade. This payment software is still in the testing phase.
app_icon
ChainCatcher Building the Web3 world with innovations.