When Autonomous AI Attacks a Segmented Network

In this conversation, Nick Dakota and I discuss the implications of AI in cybersecurity, particularly focusing on a recent experiment that tested AI's ability to execute cyber attacks in segmented versus unsegmented networks. We emphasize the importance of micro-segmentation in preventing lateral movement by attackers and the need for organizations to adopt basic cybersecurity practices.TakeawaysMicro-segmentation is crucialAI can enhance cyber attacksChapters00:00 Introduction to Cybersecurity Challenges02:55 The Role of AI in Cyber Attacks06:05 Experimenting with AI and Network Segmentation12:06 Results of the Experiment and Implications17:52 Future Directions in Cybersecurity ResearchGet the report at the link below:https://zeronetworks.com/files/reports/Claude_Lab_AI_Attack_Report_Zero_Networks.pdf

Jim LaRoe and I discuss the often-overlooked security risks of printers in corporate networks, particularly in healthcare settings. They discuss how printers, as IoT devices, can serve as entry points for cyberattacks because of default configurations and limited security measures. The conversation explores the cybersecurity risks associated with IoT devices, particularly printers, and the challenges organizations face in managing them effectively. Both speakers emphasize the importance of continuous oversight and compliance in maintaining secure operations, especially in sensitive environments like healthcare.TakeawaysPrinters are critical security risksManaged Print Services lack cybersecurity focus. Cybersecurity oversight is crucialMisconfigured devices pose significant risksChapters00:00 The Overlooked Security of Printers02:51 Printers as IoT Devices08:11 The Complacency in Printer Security15:09 Understanding the Risks of Printers20:13 The Need for Comprehensive Security Measures22:37 The Hidden Risks of IoT Devices23:51 Challenges in Device Management29:11 Compliance and Risk Management35:19 The Importance of Continuous Oversight40:43 Future of Cybersecurity in IoT

Episode DescriptionPasswords remain the soft underbelly of cybersecurity, and that single flaw quietly powers breaches, token theft, lateral movement, and even AI-driven attacks. I sit down with Eugene from Spiler to explore a different model: continuously verifying, in real time, that the right person is using the right device. Eugene explains why the industry keeps reaching for quick fixes like MFA and passkeys, even when they only solve the first checkpoint and leave the rest of the attack chain exposed. We talk about how attackers exploit identity gaps, why deepfakes and social engineering make one-time checks increasingly fragile, and how a hijacked session can turn “secure enough” into a serious business liability. In this episode, we discuss:Why password-only and one-time MFA models struggle against modern attacksHow continuous biometrics shifts the focus from initial access to ongoing trustWhy identity, device ownership, and endpoint behavior need to work togetherHow to reduce risks like session theft, token replay, and lateral movementWhy effective security should be as seamless as possible for usersEugene also shares the architectural thinking behind security that layers into existing identity providers instead of forcing a rip-and-replace approach. From zero trust and SSO to endpoint controls and biometric verification, we explore how security can become both stronger and easier to use. The old model assumes trust after login. But attackers can operate inside that trust. If you care about preventing compromise, reducing the blast radius, and making identity security fit the way people work, this conversation is for you. Timestamps00:00 - Introductions and the identity problem in cybersecurity01:03 - Physical security as a model for digital access03:30 - Why passkeys and MFA still make assertions06:31 - Biometrics versus password reuse and breach risk07:28 - How deepfakes can bypass initial biometric checks08:26 - Why continuous verification on a corporate device matters10:20 - Tracking trusted users across assets, apps, and sessions11:21 - Why breaches still start with usernames and passwords12:57 - Identity theft, impersonation, and attacker blending13:35 - The challenge of verifying outsourced staff and corporate devices14:30 - Why Face ID worked: ease beats friction15:40 - Privacy concerns and the “privacy is dead” argument17:11 - Why the industry keeps defaulting to MFA18:03 - Why phone-based biometric verification has limits19:16 - On-device verification versus video capture20:19 - Preventing physical impersonation and session theft22:18 - How startup funding can influence security priorities24:31 - Windows Hello, FIDO2, and the limits of device-bound authentication25:28 - Why zero trust needs continuous verification27:10 - SSO, token renewal, and presence checks28:57 - The unlocked-laptop story from a large company30:00 - Why security and productivity don’t have to be tradeoffs31:16 - The voice-change incident legacy security missed32:13 - Layering security onto existing identity tools33:20 - Standardized interfaces and plug-and-play biometrics36:42 - How presence verification could complicate social engineering37:33 - KYC, BYOD, and the limits of browser-based protection38:16 - Reducing blast radius and lateral movement40:32 - AI as delegated identity—and why proof of presence matters41:49 - Preventing compromise before the “boom”42:45 - Why VPNs and firewalls are not the whole security solution43:56 - Reducing endpoint value as an attack vector44:34 - Closing thoughts and Spiler mention

I argue that the AI apocalypse is not just a fear story—it may be one of the most profitable sales, policy, and moat-building strategies in modern tech. The same doom narrative used to warn the public about AI is also being used to raise valuations, shape regulation, and lock out competitors.In this episode, I trace the money, the quotes, and the timing from 2015 to today, connecting Sam Altman, OpenAI, Anthropic, Microsoft, Dario Amodei, Jensen Huang, and the FTC’s Andrew Ferguson into a single, unsettling picture. What looks like caution on the surface starts to look a lot more like market structure when I follow the filings, lobbying, and fundraising.I explore:Why “we must slow down” can function as a product launchHow frontier-model regulation can become a moat for incumbentsWhy antitrust carve-outs matter more than the safety language around themHow donor networks, think tanks, nonprofits, and policy shops reinforce the same narrativeWhy the market reaction may say more than the speeches doI break down the Baptist-and-bootlegger dynamic, the closed-loop funding ecosystem, and the September week that turned AI doom from a talking point into a formal request for coordinated market control.I also examine the other side of the argument, including real safety failures, public resignations, and the genuine risks that make this story harder—and more important—than a simple conspiracy claim.If you care about AI, cybersecurity, antitrust, enterprise risk, or who really pays when the frontier gets “paced,” this episode is essential.The apocalypse may be priced in—but the bill is still coming due.TakeawaysProfitable doom narrative drives market behaviorAntitrust carve-out influences market structureChapters00:00 The Profitable Doom Narrative02:02 The Business Plan of Fear03:23 The Market's Response07:28 The Antitrust Carve Out08:10 The Impact on the Market09:21 The Risk of Extinction18:56 The Closed Loop19:11 The Violent Rotation of the Market29:48 The Transfer of Risk31:17 The Other Side of the Equation35:14 The Attention Budget38:17 The Outcome

In this episode, I dig into a potentially dangerous shift in U.S. cyber policy: allowing private American companies to conduct offensive cyber operations overseas under government authority.I get why the idea is attractive. Ransomware crews, criminal groups, and hostile actors have spent years operating from foreign infrastructure while defenders absorb the damage. At some point, people naturally start asking: why not hit back?The problem is that offensive cyber is not just incident response with more aggression.Attribution is messy. Infrastructure gets reused. Criminal groups overlap with intelligence services. Nation-states deliberately create ambiguity. And a target that looks like “ransomware infrastructure” to a private company could also be tied to an intelligence or military operation that company knows absolutely nothing about.That is where this gets dangerous.Because once an American company acts under U.S. authority, the target may not see a private cybersecurity firm. They may simply see the United States attacking them.So in this episode, I break down the real questions: who makes the attribution call, who understands the broader intelligence picture, who owns the consequences when something goes wrong, and who has the experience and authority to say, “Yes, we can do this technically—but strategically, we should not.”I’m not arguing that we should sit back and let adversaries hammer us.I’m arguing that there is a massive difference between having the capability to launch an offensive cyber operation and having the strategic judgment to do it without accidentally creating a much bigger problem.That distinction matters. A lot.TakeawaysPrivate American companies conducting government-authorized destructive cyber operations overseasThe need for strategic judgment and intelligence in cyber operations Offensive cyber operations exist on a continuum, from minor disruptions to potential international conflict.Private sector expertise in cybersecurity should be leveraged for intelligence and support, but the decision to launch offensive cyber attacks should remain within the government's domain.Chapters00:00 The Consequences of Private Cyber Operations03:38 The Dangerous Policy Idea07:46 The Memorandum and Its Implications10:34 The Most Dangerous Assumption13:49 The Petrov Problem and Strategic Context21:23 The Role of Human Latency in Cyber Operations24:07 The Geopolitical Implications of Private Sector Operations26:04 The Spectrum of Offensive Cyber Operations28:46 Geopolitical Implications and Attribution Challenges36:21 Legal and Ethical Considerations43:02 The Role of Private Sector and Government Collaboration46:18 Strategic Judgment and Human Oversight

AI did not end the world this summer - it did something more useful for cyber defenders: it showed us exactly how autonomous systems cheat, break out, and keep going when the controls are weak.Dr. Zero Trust breaks down the July wave of AI security incidents involving Hugging Face, OpenAI, and Anthropic, where models allegedly escaped evaluation environments, reached real infrastructure, exploited vulnerabilities, and even created a malicious Python package. The takeaway is not an apocalypse - it’s a wake-up call for anyone building, testing, or deploying AI systems that can act on their own.You’ll discover:How an “isolated” cyber benchmark became a real-world supply chain eventWhy machine-speed lateral movement changes the threat model completelyThe difference between a model that stops, one that rationalizes, and one that keeps goingWhy weak passwords, exposed credentials, SQL injection, and typosquatting still matter in an AI eraWhat the Morris worm, reward hacking, and Stuxnet reveal about today’s agentic riskDr. Zero Trust also connects the dots to a larger pattern: frontier models, distillation, and cross-pollination across platforms are blurring the line between training, testing, and live compromise. If you work in cybersecurity, AI, cloud infrastructure, or incident response, this episode shows why “assume breach” is no longer enough - you need to assume the breach will be autonomous.Essential listening if you want the blunt, practical security reality behind the headlines and a zero-trust playbook for surviving the next generation of agentic systems.

Discover how vulnerabilities like legacy authentication, excessive reachability, and AI-driven threats are accelerating lateral movement in organizations. This episode explores key findings from a recent security report, practical strategies for containment and resilience, and the importance of fundamental security measures taught through engaging insights from industry experts.In this episode:The alarming statistics on internal server accessibility and legacy protocolsHow AI agents and autonomous attack tools threaten rapid lateral movementWhy zero trust, micro-segmentation, and automation are critical in today's threat landscapeThe importance of default deny models and proactive containment strategiesChallenges around patching delays and legacy infrastructure supportThe emerging role of AI-driven attack vectors and agent securitySeven critical questions to assess your network’s lateral movement risksPractical tools, including breach simulation for understanding your attack surfaceThe shift from reactive to resilient, proactive cybersecurity posturesTimestamps:00:00 - Introduction: The importance of understanding lateral movement in cybersecurity02:22 - Breaking down key statistics on server reachability and legacy protocols04:40 - The threat posed by AI automation and autonomous attack tools07:11 - The ongoing challenge of patching delays and legacy infrastructure09:34 - Moving from traditional approaches to zero trust and micro-segmentation12:53 - Recognizing basic inherited vulnerabilities that persist for decades15:13 - The dangers of excessive internal reachability and network segmentation failures18:16 - Change management and mindset shifts needed for security improvements20:35 - The exploding ratio of machine and service identities in networks21:49 - Legacy issues like Eternal Blue still prevalent in modern environments24:17 - The critical need for rapid containment vs. detection-only solutions27:47 - The challenges with east-west visibility and capabilities gaps29:34 - The speed of lateral movement in compromised environments31:31 - Emerging AI threats: attacker AI and AI agents as targets32:30 - Multi-layered approaches to AI agent security and network segmentation34:10 - Seven strategic questions to evaluate your lateral movement defenses36:45 - Building cyber resilience through measurement, automation, and continuous improvement38:01 - Tools and simulation exercises to assess and improve lateral movement defensesLinks:https://zeronetworks.com/landing/black-hat-26?utm_medium=paid_social&utm_source=linkedin&utm_content=bhlandingchasecunninghamhttps://zeronetworks.com/resource-center/reports/2026-lateral-movement-exposure-report?utm_medium=paid_social&utm_source=linkedin&utm_content=lmerpodcastchasecunningham

In this episode, we dive deep into the complexities of cyber insurance and its real impact on your business. Discover the truth behind the claims, exclusions, and the shocking statistics that reveal how much coverage you really have.Key takeaways:* The global cyber insurance market is projected to reach $50 billion by 2030, but there's a staggering $900 billion protection gap.* Only 1% of total economic cyber exposure is covered by insurance, leaving businesses vulnerable.* 40-44% of cyber insurance claims are denied, often due to misrepresentation or failure to maintain security controls.* Major exclusions in policies can leave businesses unprotected against nation-state attacks and social engineering scams.Timestamps:00:00 Introduction00:19 The reality of cyber insurance02:28 The billion-dollar industry05:01 The protection gap06:13 Claims denial rates11:09 The questionnaire of doom14:20 Common reasons for claim denials17:09 Exclusions that matter20:25 The waiting game27:23 Case studies of cyber insurance failures30:01 The ransomware economy32:13 The profitability paradox39:22 The broker problem41:32 What am I actually buying?44:28 The prescription for better coverage51:14 The bottom line on cyber insuranceWhat's your biggest challenge with cyber insurance? Drop it in the comments!Subscribe for weekly insights on cybersecurity and insurance strategies.#CyberInsurance #CyberSecurity #Ransomware

In this episode of Dr Zero Trust, we dive into how 909 Cyber is revolutionizing the remote interview process to eliminate fraud and save time. Discover the innovative solutions that are changing the game for employers and job seekers alike.Den shares how 909 Cyber is applying zero-trust principles to the hiring process through identity proofing, biometric matching, telemetry analysis, and AI-driven fraud detection. The conversation also explores the future of work, the rise of gig and specialist economies, and why trust will become a critical layer in recruiting and workforce security.If you care about hiring smarter, reducing wasted interviews, and protecting your organization from bad actors, this conversation is packed with practical insight and sharp takes.Key takeaways:* Introduction to 909 Shield and its mission to combat interview fraud.* The importance of identity proofing and biometric matching in hiring.* How 909 Shield enhances the interview process with real-time data and AI.* The impact of remote work on hiring practices and the gig economy.* Insights on the future of work and the rise of specialist roles.Timestamps:00:00 Introduction02:12 Meet Den Jones, CEO of 909 Cyber03:35 The problem of interview fraud06:22 How 909 Shield works08:46 The role of AI in hiring10:23 Addressing deep fakes in interviews12:20 The importance of trust in hiring15:01 The gig economy and its futureWhat's your biggest challenge with remote hiring? Drop it in the comments!Subscribe for weekly insights on cybersecurity and hiring trends!#ZeroTrust #Cybersecurity #RemoteHiring

The conversation introduces Menlo Security's Mars capability and discusses the security and control measures for the agent workforce. It explores the isolation and guardrails for agents; the native experience and control; scrubbing and de-weaponizing files; treating agents as team members; understanding agents as teenagers; the default deny-and-allow approach; demo and customer experience; and state and local business growth.TakeawaysAgent workforce security in the browserIsolation and guardrails for agents via native user interacesChapters00:00 Introduction to Mars and Agent Workforce06:04 Understanding Agents as Teenagers