Podcast Summary

To The Point - Cybersecurity

Episode: The Merging Worlds of AI, Cybersecurity, and Physical Threats with David Saunders

Date: December 23, 2025
Host(s): Rachael Lyon & Jonathan Knepher
Guest: David Saunders, Director of Forcepoint Security Lab

Overview

This episode explores the increasing convergence of artificial intelligence (AI), cybersecurity, and physical threats. Rachael and Jonathan are joined by David Saunders, a veteran cyber threat researcher, who shares insights on the latest threat patterns, the rapid evolution of AI, and its impact on both cyber and physical security domains. The discussion delves into modern attack methods, the vulnerabilities of backup systems, AI-enabled phishing, and the challenges facing conventional security technologies.

Key Discussion Points & Insights

1. The Convergence of AI, Cybersecurity, and Physical Threats

• Main Point: The convergence of AI with cyber and physical security domains is accelerating, leading to new, complex threat landscapes.
• Saunders observes:
  • There’s a “race to AI-fi everything” across industries, often with security and safeguards as an afterthought. (03:06)
  • Increased reporting on attacks affecting both cyber and physical systems; the UK's National Cybersecurity Center reports a threefold increase in such incidents.
  • “Even cybersecurity companies are challenged with the speed at which AI has come about, and we're supposedly the experts in that space.” (04:22, David Saunders)

2. Notable Threat Patterns in 2025 and Beyond

• Patterns Observed:
  • Lateral movement and privilege escalation in networks remain prevalent attack methods. (05:39)
  • Backup systems have become high-value targets, with attackers aiming to both encrypt primary data and compromise backup copies to increase ransom leverage. (07:00)
  • The media increasingly reports on longer system downtimes post-attack, reflecting the complexity of modern recovery efforts.
  • “People are still our weakest link…AI is very adept at trying to even push those boundaries of duping people.” (09:44, David Saunders)

3. The Persistent Problem of Backup Vulnerabilities

• Discussion:
  • Modern backups are typically online, making them susceptible to ransomware.
  • Saunders recommends a combination of online (cloud) and offline (air-gapped or removable) backups to prevent single-point failures:

    “Maybe again multiple locations, most organizations probably thinking about that as well, but certainly maybe not having it all online...having some sort of air gap or physical disconnection from it.” (11:18, David Saunders)

  • The group jokes about reverting to “the Stone Age” as an extreme, referencing the possibility of shutting down connectivity for critical infrastructure. (11:36-11:54)

4. AI’s Impact on Content Quality and Web Security Policies

• AI-generated Content:
  • Explosion of low-quality, AI-generated sites ("AI garbage") complicates web filtering and content validation. (14:04-14:30)
  • Jonathan and David riff on the idea of a new filter category: “AI spam” for websites that are only machine-generated filler.

    “Maybe Arsenal needs a new category which is effectively like a park domain, but for AI rubbish…AI spam, if you want to call it in a web context.” (15:05, David Saunders)

5. Sophisticated Phishing and Social Engineering with AI & LLMs

• Emergence of AI-powered Phishing:

  • The end of typo-laden phishing:

    “The era of typos is over...” (17:00, David Saunders)

  • Advanced LLMs allow attackers to personalize content with scraped social and business information for highly believable attacks.
  • The ability for AI—especially in audio attacks—to adapt in real time, shifting strategies based on user hesitance or skepticism.
  • The overall trend is toward increased efficiency and scale, not necessarily new attack vectors.

• Philosophy of Defense:

  • The security balance: high-value targets justify costly, layered defenses, but AI reduces the cost and increases attackers’ effectiveness, eroding this assumption.

    “If it makes things cheaper, it enables them to more effectively attack either more candidates or even go after bigger candidates.” (20:57, David Saunders)

6. AI’s Role in Malware Evolution and Security Industry Shifts

• Malware Mutation & Detection Challenges:
  • AI can easily produce endless polymorphic malware variations, overwhelming signature-based detection and forcing a transition to heuristic-based approaches. (24:21-25:30)
  • Potential resurgence of sandboxing technologies:

    “Maybe, you know, sandboxing could be a spin off, or a beneficiary, if you like, of some of this AV variations.” (25:55, David Saunders)

Notable Quotes & Memorable Moments

• “There's a definite sort of race to AI-fi everything...security being a secondary thought and safeguards aren't there, put in place where they should be.” (03:09, David Saunders)
• “UK’s National Cybersecurity Center reported a threefold increase between cyber and physical attacks being reported.” (03:36, David Saunders)
• “The era of typos is over.” (17:00, David Saunders)
• “Maybe Arsenal needs a new category which is effectively like a park domain, but for AI rubbish…AI spam, if you want to call it in a web context.” (15:05, David Saunders)
• “It's all about being more efficient, faster, and actually probably the most important thing is cheaper.” (19:14, David Saunders)
• “You've sufficiently caused a lot of alarm.” (26:41, Jonathan Knepher)

Timestamps for Key Segments

• [03:06] – The challenge of AI speed outpacing security safeguards
• [05:39] – Common threat patterns: lateral movement, privilege escalation, backups
• [09:44] – AI’s power to exploit human weaknesses
• [11:18] – Advice on backup isolation and air-gapping
• [14:30] – AI-generated content and the need for new spam filters
• [17:00] – AI-powered phishing and the “end of typos”
• [20:57] – Security economics disrupted by AI efficiency
• [24:21] – AI-fueled malware evolution and detection arms race
• [25:55] – Sandboxing as a potential solution to AI-augmented malware

Tone and Language

The conversation is energetic but earnest, balancing technical depth with relatable analogies and a touch of humor. Saunders brings a pragmatic, sometimes cautionary, perspective, while Rachael and Jonathan infuse warmth and accessibility.

Conclusion

This episode paints a vivid picture of the challenges facing cybersecurity in a world rapidly shaped by AI. The merging of digital and physical threats, coupled with the democratization of attack tools via AI, demands new strategies, smarter defenses, and a rethinking of longstanding practices. The hosts emphasize vigilance, layered defense, and continuous adaptation as the only viable way forward.