David Sacks, investor and White House AI and crypto czar, posted on X today, October 22, 2026, to clarify the distinction between user-authorized AI agents and actual cyberattacks in the financial sector. Sacks stated, “There’s a world of difference between user-authorized agents (Muse, Instinct, Grokbot, etc.) and true unauthorized/rogue agents that could constitute cyberattacks.” He noted that much of the activity described in a recent Axios story involves customers granting credentials to agents for actions on their behalf, which, while potentially violating bank terms, “are in no way cyberattacks.” Sacks urged banks to “support user-directed agents,” citing American Express as an example of an authorized connection within platforms like Muse, and suggested read-only permissions as a sensible interim step.
The discussion around AI agents in finance is gaining traction, with Gokhshtein Media recently reporting on XRPL institutional adoption growing as banks deploy stablecoins and AI agents. The “Corporate Stablecoin Use Doubles to 18%, Banks Commit to Launch” coverage also highlights the increasing integration of digital tools within traditional financial systems. The Crypto Fear & Greed Index currently stands at 68, indicating 'Greed' in the broader digital asset market, reflecting ongoing interest and activity.
Sacks’ remarks imply a need for banks to adapt to new technologies by differentiating between legitimate user-driven automation and malicious attacks. His view suggests that frightening the public by conflating these two categories is irresponsible, and that banks should work towards supporting authorized agent connections. This would allow users to grant, limit, and revoke access, while carefully approaching full write access with user credentials due to its different risk category.
