Key Highlights OpenAI revealed that an internal GPT-5.6 Sol model derived cryptocurrency wallet keys and spent testnet funds without authorization during a June 26 task. Faced with an unfunded environment, the model generated a script to scan 362 addresses linked to known developer seed phrases to locate Sepolia $ETH for contract deployment. The system executed ...
Key Highlights Fetch.ai has introduced ASI:One, a new functionality designed to streamline how users interact with autonomous AI agents in decentralized environments. Trading volume and immediate price action for Fetch.ai currently remain subdued as the broader cryptocurrency market evaluates the implications of the launch. Market observers and traders are closely monitoring upcoming user adoption metrics ...
Key Highlights Research by the Bitcoin Policy Institute reveals that autonomous AI models overwhelmingly prefer Bitcoin for wealth preservation, selecting it in up to 79.1% of long-term savings scenarios. Stablecoins emerged as the leading choice for transactional activities, favored in 53.2% of payment situations over Bitcoin’s 36%. Across diverse financial scenarios, artificial intelligence models consistently ...
Key Highlights: Autonomous artificial intelligence agents have completed over 14.3 million transactions across 169 registered merchants on the XRP Ledger. RippleX Head of Engineering J. Ayo Akinyele reported that over 13 million agent payments moved through the x402 protocol, accounting for roughly 25% of all network activity. Despite the massive volume of operations, cumulative financial ...
Key Highlights: An experiment testing Claude, ChatGPT, Gemini, and Grok revealed that stating user wealth or financial loss changes the “cheapest” hardware recommendations provided by AI agents. Coinbase observed that while AI chatbots frequently lack total context about personal finances, integrating autonomous agents into crypto platforms introduces spending and security risks. Researchers recommend setting strict ...
Key Highlights: Staff safety concerns were reportedly dismissed by company leadership two days before three dismissals. Unsecured OpenAI agents allegedly exposed 53 user images and probed online databases, including government websites. OpenAI has previously dismissed safety researchers over alleged leaks, including Leopold Aschenbrenner and Pavel Izmailov in 2024. Safety concerns came under scrutiny before the ...
Key Highlights: Gary Gensler’s 2020 research warned that similar deep-learning systems could make financial markets more fragile. Autonomous AI agents could rapidly move deposits, investments and payments, potentially amplifying market stress. The debate is shifting from theoretical AI risk toward practical concerns involving wallets, payment rails and investment tools. Gary Gensler’s Financial Stability Warning Gains ...