Elon Musk predicts that super-intelligent AI agents could exploit financial system vulnerabilities to make money for themselves, instead of serving humans.

In a new comment on his X page, the American technology billionaire wrote: “At some point, superintelligent agents will do the same thing to make money for themselves, not for humans.”

The statement came after hackers used crypto assets stolen from Binance as collateral on decentralized finance platform Aave to borrow Ether. According to LlamaRisk analysis, one wallet involved in the attack deposited 60 million USDT into Aave V3 and borrowed 19,762 WETH, worth approximately $44.6 million at the time of the transaction.

According to Bernardo Brites, CEO of Trace Finance, AI tools are now capable of finding security vulnerabilities faster than engineering teams can patch them. He believes that this difference in speed makes the risk of large-scale attacks increasingly concerning.

Not only Musk, technology and Bitcoin investor Chamath Palihapitiya also warned about the risk of Bitcoin being attacked in the future, especially as quantum computer technology and AI's self-improvement capabilities continue to develop. He even mentioned the scenario that Bitcoin could lose its value if the security system is broken.

However, investors still have more cautious assessments. Yan Pritzker, Swan Bitcoin application technology director, said that Binance's commitment to compensate for all losses helps limit the risk of the crisis spreading. He also said that the Bitcoin market currently has liquidity and the participation of large organizations, helping to increase its resilience to shocks.

The Binance attack caused the Fear and Greed index - a measure of cryptocurrency market sentiment, to move into the "fear" zone. Even so, some investors still believe Bitcoin can recover from the volatility.

Although he did not confirm that AI was involved in the incident, Musk's comments suggest an interesting scenario in cybersecurity: As AI agents become increasingly capable of operating on their own, they may seek to exploit technological and financial vulnerabilities to achieve established or self-formed goals.

Autonomous AI capabilities also require mechanisms for monitoring, limiting access, and controlling behavior, especially when these systems are empowered to interact with valuable accounts, data, and assets.