Crypto leaders warn AI agents could scale financial hacks and complicate liability
Panelists at the Wyoming Blockchain Symposium warned that while artificial intelligence agents reduce friction in decentralized finance, they simultaneously introduce severe security, privacy and legal liability risks for investors.
Industry executives warned at the Wyoming Blockchain Symposium 2026 that autonomous artificial intelligence agents could dramatically scale financial cyberattacks across decentralized networks. While these tools promise to eliminate friction in crypto navigation, leaders argue they simultaneously lower the barrier to entry for malicious actors targeting digital wallets and personal data.
Global Settlement Network Chief Executive Ryan Kirkley cautioned that current billion-dollar bridge hacks are merely "pennies" compared to the potential scale of agentic threats. He noted that while targeting a single individual with $20,000 in assets was previously too resource-intensive for hackers, autonomous agents can now systematically pursue every user. "Now I can have an agent that goes after everyone," Kirkley said.
Web3 Foundation Vice President of Technical Operations Bill Laboon agreed that the efficiencies benefiting decentralized systems cut both ways, creating less friction for bad actors alongside good ones. Laboon also highlighted severe privacy vulnerabilities, warning that artificial intelligence could exploit metadata leaks to reconstruct supposedly secure transactions on private chains.
The delegation of authority to these systems remains a critical vulnerability. Midnight Foundation President Fahmi Syed argued that granting a single agent unrestricted access to credit cards, Social Security numbers and multiple financial accounts is a "scary concept" without strict operational parameters.
Kirkley viewed setting agent authority as a secondary issue compared to securing the underlying infrastructure. He questioned whether the industry is simply "creating an attack vector where an agent could be taken over and drain an entire wallet."
Beyond security, trust and model reliability pose significant hurdles to mainstream institutional and retail adoption. Laboon pointed to the persistent tendency of large language models to hallucinate as a primary reason investors should hesitate. "I wouldn't want to put my 401k in the hands of that," he said.
Silvermine Capital Advisors founder Richard Shorten observed that agentic technology has "moved further and faster" than the market can comfortably process. He noted that the central challenge is translating raw capability into a trusted toolkit for business and personal finance.
Regulatory frameworks and legal liability present the final major bottleneck for autonomous financial agents. Kirkley emphasized the difficulty of determining accountability when an independent system breaks the law or misallocates funds. "If your agent goes and does something illegal, or does something where it spends money that you can't have, how do you claw it back?" Kirkley asked, highlighting the unresolved question of where the ultimate legal buck stops.
