Infosecurity Europe 2026, one of the leading cybersecurity conferences in the UK, centred on cyber resilience, proactive threat defence and security risks brought by Agentic AI this year.
The three-day event took place from June 2 to June 4, 2026, organised by RX Global in Excel London.
The conference brought together a distinguished number of individual experts and enterprises, including Jason Fox, who served in the UK Special Forces and operated in some of the world’s most high-risk environments. He delivered his keynote on Thursday 4 June, 10:05–10:45, on Decision-Making Under Pressure: Leading Through Live Cyber Incidents.
Other features included a former Special Boat Service Sergeant and SAS: Who Dares Wins star. The line-up spanned cybersecurity, law enforcement, elite sport and the military, united around the themes of leadership, resilience and innovation in the modern cybersecurity enterprise space.
As a key media partner of Infosecurity Europe 2026, EM360Tech was on the ground at one of the cybersecurity industry's most anticipated events of the year.
We caught sight of several experts for an inside scoop on cybersecurity and what’s going to be the most prevalent theme dominating the industry next year.
Agentic AI in cyber resilience was among the most talked about at the event this year.
Key Challenges at Infosecurity Europe 2026
The experts at Infosecurity Europe 2026 addressed the key challenges of emerging technologies.
In a vox pop interview at the event floor, Ariel Zamir, CEO of Ray Security, said to Dua that enterprises approach AI security in an infeasible way. They try to use the same solutions to address identical problems.
“Everybody thinks about AI in the same way,” he said. “Every enterprise has the same type of problem, but all the solutions around address the problem the same way.”
“That is infeasible, and this is the key challenge in cybersecurity.”
At Infosec, Mike Britton, Chief Information Officer, Abnormal AI, told Dua that his panel discussion spotlighted Anthropic Mythos, AI, and the concerns that keep everyone up at night.
The biggest challenge, he said, is the pace of innovation and how fast things are moving.
“You used to have 18 months to catch up, and now you might have 18 days.” He says that that's a challenge for all enterprises. “Attackers are taking advantage of this pace of innovation, and defenders are having a tough time staying on top of it.”
When AI Spend Demands Proof
How enterprises are shifting from pilots to disciplined AI value management that ties every use case to outcomes finance leaders trust.
Cyber Resilience is Key
Meanwhile, Jon Fielding, Managing Director, Apricorn, says that the price of memory is now going through the roof and “we’re trying to keep pace with that and ensure our pricing is not out of alignment with the AI increases.”
Steven Peake, Director, Solutions Architects, Northern Europe at Barracuda, says that the focus is moving away from just protection. “They’re looking at capabilities that detect and respond when the protection layers are missed. They’re ensuring that they still have recoverability in terms of backup and data recoverability.”
“That's what we call cyber resilience,” he added.
Adam Boynton, Senior Security Strategy Manager, EMEIA, Jamf, addressed cybersecurity challenges for Mac devices particularly.
Challenges CISOs face have been pertinent to bringing Mac into the enterprise, which has been growing in use, but it hasn't had the same amount of control or protection that a Windows environment has in an enterprise space.
“A lot of existing security tools are very Windows-focused,” Boynton stated. “Jamf is focused on the Macverse ecosystem and now is a mobile offering too.”
Inside Nvidia’s RTX Spark PCs
A look at how Arm-based RTX Spark, unified memory and on-device LLMs reshape Windows hardware design for AI-intensive workflows.
“Wannabe Hackers” More Dangerous Than “Hardened Criminals”
Addressing one of the biggest challenges, Amanda Finch, CEO, Chartered Institute of Information Security (CIISec), tells Dua that the wannabe hackers are more dangerous than hardened criminals.
These "wannabe hackers" with limited training can be particularly dangerous because they may apply AI tools without understanding the consequences or acting responsibly.
“You’ve got people without much training on the dark side,” she added. “They don’t know what they’re doing when they’re using these new technologies. They don’t apply it responsibly.”
Mental Fatigue & Scale of AI Innovation
Merlin Gillespie, Director, Cybanetix, told Dua that there’s mental fatigue in the industry.
“We’re facing a generic challenge – nobody is entirely sure what the problems are going to look like past three to four months from today.”
He further added that cybersecurity faces a combination of problems today. There is a lack of certainty regarding future threats, as professionals find it difficult to predict what problems will look like beyond a few months.
Steve Povolny, Vice President of AI Strategy & Security Research, Exabeam stressed that the current industry landscape is characterised by the rapid speed and scale at which AI and autonomous agents are being deployed, presenting both significant threats and new opportunities.
“The speed and scale of AI, autonomous agents and AI Agents in general are one of the biggest threats and opportunities we have had in the cybersecurity industry in the last 20 years.”
Why Frontier AI Stays Locked
By limiting access to Mythos, Anthropic highlights the governance gap between AI’s cybersecurity potential and the systemic risk of weaponized models.
Positive Outlooks on Cybersecurity
James Savory, Regional Vice President - UK & Ireland, at Island echoed a fundamental question teams are grappling with – how to enable the adoption of AI without introducing new vulnerabilities.
“Because the browser is identity-aware, context-aware, and intent-aware, married with the controls Island puts in enterprises can govern what happens there,” Savory tells Dua. “We enable users to safely adopt AI while stopping the bad things.”
On the other side, Rik Ferguson, Vice President, Security Intelligence, Forescout, proposed a new paradigm – “Assume Autonomy.”
Enterprises and threat actors are increasingly building autonomy into their systems. There is documented evidence, including reports from Google and Anthropic, of attacks being 80-90 per cent autonomous with human guidance.
“AI has been a buzzword at events for a couple of years now. All the marketing messaging revolves around AI, but it’s important to remember that AI has been engineered for speed, not accuracy,” Ferguson tells Dua.
Inside Nvidia’s Korea AI Bet
Nvidia locks in Korean chip, cloud and robotics partners to anchor its global AI build-out and secure both supply and demand.
Assume Autonomy
Ferguson’s Assumed Autonomy means assuming autonomy within the defensive stack but with the right guardrails in place.
“We also need to assume autonomy from the attacker because they will be building AI and autonomous systems into their attack frameworks.” This means they’re not operating at human speed anymore or within a human mental paradigm. All frameworks are typically built on the assumption that the adversary is human.
Supply Chain Security & AI
In another interview with Ofri Ouzan, Security Researcher at JFrog, she explains the trends in supply chain security, the impact of AI on development, and how to manage the resulting surge in vulnerabilities.
From a security perspective, in terms of how AI helped find more vulnerabilities, JFrog witnessed a significant increase in the number of existing vulnerabilities. “2025 had a 20 per cent upgrade from 40k to 50k vulnerabilities,” Ouzan tells Dua. “Some vulnerabilities can be false positives.”
“Criminal Underworld” Insights
However, Thom Langford, EMEA CTO at Rapid7, highlighted the shift toward preemptive measures and the impact of artificial intelligence (AI) on cybersecurity enterprises.
“We as an industry have moved away from detection and response being the solution to everything. Now, we’re looking at the other side, which is the preemptive side.”
“We’re actually trying to stop threats in the first place.”
It means understanding the attack surface, and understanding what the attackers are seeing using threat intelligence to preemptively make decisions based on what’s actually happening in the “criminal underworld.”
In the past, threat intelligence didn’t have the kind of fidelity, quality or timeliness like right now.
Crisis Management in AI Cybersecurity
In the interview with Daniel Lattimer, Vice President, EMEA, Semperis says that identity resilience is not the sole solution.
“We’re not only looking at identity resilience. We’re prepared to respond to attacks and progress, but also crisis management.”
However, Jon Abbott, CEO, ThreatAware, spotlighted that the industry often focuses too much on blaming the user or solely on phishing, which is considered too simplistic.
Instead, Abbott makes a case for a return to fundamentals—such as utilising web proxies and EDR everywhere—to prevent various types of attacks, including those powered by AI.
“You really got to look back at those controls and appreciate you could have a web proxy or make sure you have Endpoint Detection and Response (EDR) everywhere instead of blaming the user for clicking links.”
AI Regulations During Developmental Phase
Joseph Byrne, Solutions Engineer, Vanta, talked about the impact of AI regulations during an exciting technological developmental period in cybersecurity.
“We’re trying to balance the innovation with compliance,” Byrne tells Dua. “I’m excited about finding the balance with them for greater things, but do them with governance.”
Quantum Computing: The Future of Cybersecurity
In another vox pop at Infosec 2026, Jake Moore, Global Cybersecurity Advisor, ESET, said to Dua that the industry was 10 years away from quantum computing, and now he believes we’re three to five years away.
When asked about enterprises' reluctance to adopt cryptography, Moore says that top executives are misunderstanding the technology and perceiving the initial AI hype versus threat.
“There’s a knowledge gap,” he added. “Looking at quantum, we don’t fully know the technology yet, but we need to jump into it now.”
He says that it’s important, especially when it comes to breaking encryption. “If we’re seeing criminals harvest all of that data knowing that now they can decrypt in the future, we’re sitting on a time bomb that’s going to up in the next five or 10 years.”
Comments ( 0 )