AI, Cybersecurity and Compliance In the Agentic Era Panel - 20 October 2026, 6:30 PM

#1758341: AI, Cybersecurity and Compliance In the Agentic Era Panel - 20 October 2026, 6:30 PM - 9:00 PM

Description:	The AI Collective Bucharest with SoCyber and IPSTSO International Police Security Technology System Organization for a timely conversation on the evolving cybersecurity landscape I’ll be a panellist with some amazingly cool people in Bucharest, Romania. On Tuesday, 20 October 2026, 6:30 PM - 9:00 PM (Bucharest local time). The panellists are a group of brilliant folks who love securing AI and the future. AI, Cybersecurity and Compliance In the Agentic Era When AI can act autonomously, how do we keep organisations secure - and accountable? Agentic AI is changing the threat landscape, giving both attackers and defenders increasingly powerful tools. We’ll explore how the cybersecurity industry is adapting, how security professionals can stay ahead of emerging threats, and what the EU’s regulatory framework, including the AI Act, NIS2, DORA and CRA, requires from organisations. Speakers: Krasimir 'Krassy' Kotsev — Founder & CEO, SoCyber and Kikimora: Krasimir specialises in vulnerability management and system hardening, helping organisations strengthen their defences as they adopt AI agents. Chris Kubecka — Founder & CEO, HypaSec: Chris brings expertise in nation-state cyber incidents, ethical hacking and critical infrastructure, including helping restore Saudi Aramco’s operations after the Shamoon attacks. Toma Cimpeanu — CEO, ANSSI: A former Romanian representative on ENISA’s board, Toma brings expertise in cybersecurity governance, regulation and public-private cooperation. Cristian Ganea — Spokesperson, IPSTSO: Cristian promotes international cooperation in security, prevention and safety, connecting public institutions, civil society and the business community. Moderator: Matei Ștefan, founder Founder @ SenecAI Consulting Panel Questions AI, Cybersecurity and Compliance in the Agentic Era — Moderator Questions Opening — all speakers “With the adoption of AI, increasingly sophisticated tools are being used by both cyber attackers and security professionals. How has this changed the cybersecurity landscape, and how does it affect the day-to-day work of a cybersecurity professional?” Chris — AI changing both attack and defence “Let’s discuss international cyber-security. Poor technical security ties into geopolitical risk. You have been a key player when it comes to defending against cybercriminal syndicates. What are their strategies? How do they rely on AI tools to commit illegal acts? And how do you outsmart them? Tell us a bit about your most interesting cases as a cybersecurity expert. How did you approach the resolution of the Saudi Aramco hack? Or the attacks at the Saudi embassy in The Hague? And what lessons did the victims take from that attack? Agentic follow-up: “If an AI agent uses that information to approve a supplier or initiate a payment, how do we stop manufactured credibility from becoming an authorised transaction?” Cristian — reporting scams and making reports useful “Cristian, someone receives a fraudulent job offer or a fake work-permit offer involving several countries. What should a workable reporting process look like—from that first report to prevention and enforcement?” Follow-up: “Where does that process most often break down between the victim, the platform, public institutions and police?” Krasimir — turning 10,000 findings into decisions “Krassy, an organisation receives 10,000 vulnerability findings. How do you turn that into a defensible decision about what to fix first?” Follow-up: “What do you need to know about its assets, exposure and business operations before a severity score becomes meaningful?” Agentic follow-up: “Does a vulnerability become more urgent if an AI agent can reach it using privileged credentials?” Krasimir — cloud security and autonomous access “You’ve emphasised that being in the cloud does not automatically mean being secure. When organisations add AI agents to their cloud environments, which assumptions become especially dangerous?” Follow-up: “How should they control what an agent can access, change or authorise—and limit the damage if it makes a mistake or is manipulated?” Krasimir — using AI to reduce actual risk “Where can AI and behaviour analysis genuinely help organisations reduce risk, rather than simply generate more alerts?” Follow-up: “What evidence would convince you that an AI-driven mitigation is working—and when would you allow it to act automatically?” Toma — making regulation operational “Toma, organisations hear about the AI Act, NIS2, DORA and the Cyber Resilience Act, but may struggle to understand which obligations apply to them. How should they translate that regulatory picture into clear ownership and practical security decisions?” Follow-up: “Where do you see the greatest gap between demonstrating compliance and being able to withstand an incident?” Toma — accountability when an agent acts “If an AI agent exposes sensitive data, changes a security control or authorises a fraudulent transaction, what should an organisation already have in place to establish what happened and who was responsible?” Follow-up: “What records and oversight would make that accountability meaningful when several suppliers and systems are involved?” Cross-panel scenario — from deception to incident “Let’s bring this together. An organisation’s AI procurement agent finds a convincing supplier online, accepts fabricated credentials and initiates a payment. The supplier’s apparent credibility comes from a network of planted websites and AI-generated content. Where should this have been stopped?” Chris: “How would you investigate the deception and establish which evidence is trustworthy?” Krasimir: “Which technical controls would have limited the agent’s actions?” Toma: “How should responsibility be assigned across the organisation and its providers?” Cross-panel — sharing intelligence at machine speed “Attackers can automate across borders, while defenders still navigate institutional and legal boundaries. What needs to change so that useful threat information reaches the people who can act on it quickly enough?” Follow-up: “What can public-private cooperation realistically deliver now—and what still needs policy or legislative change?” Closing — all speakers “What is one concrete action an organisation should take in the next 30 days before giving AI agents greater autonomy?”
More info:	https://secevangelism.substack.com/p/ai-cybersecurity-and-compliance-in

Description:

The AI Collective Bucharest with SoCyber and IPSTSO International Police Security Technology System Organization for a timely conversation on the evolving cybersecurity landscape
I’ll be a panellist with some amazingly cool people in Bucharest, Romania. On Tuesday, 20 October 2026, 6:30 PM - 9:00 PM (Bucharest local time). The panellists are a group of brilliant folks who love securing AI and the future.

AI, Cybersecurity and Compliance In the Agentic Era
When AI can act autonomously, how do we keep organisations secure - and accountable?

Agentic AI is changing the threat landscape, giving both attackers and defenders increasingly powerful tools. We’ll explore how the cybersecurity industry is adapting, how security professionals can stay ahead of emerging threats, and what the EU’s regulatory framework, including the AI Act, NIS2, DORA and CRA, requires from organisations.

Speakers:

Krasimir 'Krassy' Kotsev — Founder & CEO, SoCyber and Kikimora: Krasimir specialises in vulnerability management and system hardening, helping organisations strengthen their defences as they adopt AI agents.

Chris Kubecka — Founder & CEO, HypaSec: Chris brings expertise in nation-state cyber incidents, ethical hacking and critical infrastructure, including helping restore Saudi Aramco’s operations after the Shamoon attacks.

Toma Cimpeanu — CEO, ANSSI: A former Romanian representative on ENISA’s board, Toma brings expertise in cybersecurity governance, regulation and public-private cooperation.

Cristian Ganea — Spokesperson, IPSTSO: Cristian promotes international cooperation in security, prevention and safety, connecting public institutions, civil society and the business community.

Moderator: Matei Ștefan, founder Founder @ SenecAI Consulting

Panel Questions
AI, Cybersecurity and Compliance in the Agentic Era — Moderator Questions

Opening — all speakers

“With the adoption of AI, increasingly sophisticated tools are being used by both cyber attackers and security professionals. How has this changed the cybersecurity landscape, and how does it affect the day-to-day work of a cybersecurity professional?”

Chris — AI changing both attack and defence

“Let’s discuss international cyber-security. Poor technical security ties into geopolitical risk. You have been a key player when it comes to defending against cybercriminal syndicates. What are their strategies? How do they rely on AI tools to commit illegal acts? And how do you outsmart them?

Tell us a bit about your most interesting cases as a cybersecurity expert. How did you approach the resolution of the Saudi Aramco hack? Or the attacks at the Saudi embassy in The Hague? And what lessons did the victims take from that attack?

Agentic follow-up: “If an AI agent uses that information to approve a supplier or initiate a payment, how do we stop manufactured credibility from becoming an authorised transaction?”

Cristian — reporting scams and making reports useful

“Cristian, someone receives a fraudulent job offer or a fake work-permit offer involving several countries. What should a workable reporting process look like—from that first report to prevention and enforcement?”

Follow-up: “Where does that process most often break down between the victim, the platform, public institutions and police?”

Krasimir — turning 10,000 findings into decisions

“Krassy, an organisation receives 10,000 vulnerability findings. How do you turn that into a defensible decision about what to fix first?”

Follow-up: “What do you need to know about its assets, exposure and business operations before a severity score becomes meaningful?”

Agentic follow-up: “Does a vulnerability become more urgent if an AI agent can reach it using privileged credentials?”

Krasimir — cloud security and autonomous access

“You’ve emphasised that being in the cloud does not automatically mean being secure. When organisations add AI agents to their cloud environments, which assumptions become especially dangerous?”

Follow-up: “How should they control what an agent can access, change or authorise—and limit the damage if it makes a mistake or is manipulated?”

Krasimir — using AI to reduce actual risk

“Where can AI and behaviour analysis genuinely help organisations reduce risk, rather than simply generate more alerts?”

Follow-up: “What evidence would convince you that an AI-driven mitigation is working—and when would you allow it to act automatically?”

Toma — making regulation operational

“Toma, organisations hear about the AI Act, NIS2, DORA and the Cyber Resilience Act, but may struggle to understand which obligations apply to them. How should they translate that regulatory picture into clear ownership and practical security decisions?”

Follow-up: “Where do you see the greatest gap between demonstrating compliance and being able to withstand an incident?”

Toma — accountability when an agent acts

“If an AI agent exposes sensitive data, changes a security control or authorises a fraudulent transaction, what should an organisation already have in place to establish what happened and who was responsible?”

Follow-up: “What records and oversight would make that accountability meaningful when several suppliers and systems are involved?”

Cross-panel scenario — from deception to incident

“Let’s bring this together. An organisation’s AI procurement agent finds a convincing supplier online, accepts fabricated credentials and initiates a payment. The supplier’s apparent credibility comes from a network of planted websites and AI-generated content. Where should this have been stopped?”

Chris: “How would you investigate the deception and establish which evidence is trustworthy?”

Krasimir: “Which technical controls would have limited the agent’s actions?”

Toma: “How should responsibility be assigned across the organisation and its providers?”

Cross-panel — sharing intelligence at machine speed

“Attackers can automate across borders, while defenders still navigate institutional and legal boundaries. What needs to change so that useful threat information reaches the people who can act on it quickly enough?”

Follow-up: “What can public-private cooperation realistically deliver now—and what still needs policy or legislative change?”

Closing — all speakers

“What is one concrete action an organisation should take in the next 30 days before giving AI agents greater autonomy?”

More info:

https://secevangelism.substack.com/p/ai-cybersecurity-and-compliance-in

Date added	Oct. 20, 2026, 2:23 p.m.
Source	secevangelism
Subjects	Agentic AI AI/ML - Artificial Intelligence / Machine Learning / GenAI / Artificial General Intelligence - AGI - Various Cloud Security Compliance regulations / Ethics - Various Conferences and Seminars Europe DORA - Digital Operational Resilience Act - EU EU AI Act / EU Artificial Intelligence Act EU Cyber Resilience Act Generative AI / Generative Artificial Intelligence / GenAI / GAI NIS2 - NIS Directive / Network Information Security 2 / NIS2 / EU NIS 2 / NIB2 / EU Directive on Security of Network and Information Systems. Saudi Aramco Shamoon /Disttrack Malware Vulnerability Management - Various
Venue	Bucharest, Oct. 20, 2026, midnight - Oct. 20, 2026, midnight
Country	Romania