23/10/2026
What happens when AI agents start making payments on our behalf? And who is responsible when something goes wrong?
Krasimir Kotsev explored these questions yesterday at the Bucharest Cybersecurity Conference 2026 in his session “Who Authorised the Bot?”
As AI agents move from assisting people to deciding and paying on their own, a new set of security questions follows:
🤖 Identity: who is behind the agent, and what is it allowed to do?
🎭 Manipulation: what if malicious content steers its decisions?
💱 Financial control: how do we keep it within set spending limits?
🤝 Accountability: who answers when a transaction goes wrong?
The key takeaway: AI can make the decision, but security controls must decide what it is allowed to execute.
A sincere thank you to the Romanian National Cyber Security Directorate for the invitation and for bringing together such a strong community of policymakers, industry leaders and practitioners in Bucharest.
Thank you as well to everyone who joined the session and kept the conversation going afterwards.
See you at BCC next year!