Show Notes
In this episode of BHIS Presents: AI Security Ops, the team tackles a deceptively simple question with some very complicated answers: Can you actually ban an AI model? Not access to an API. Not the chips used to train it. The model weights themselves — files that can be downloaded, copied, modified, quantized, fine-tuned, and redistributed around the world. As governments consider restrictions on Chinese open-weight models, the security argument cuts in both directions. There are legitimate concerns around national security, guardrails, model capabilities, and foreign technology dependence. But those same open models are inexpensive, locally deployable, and can give defenders capabilities that commercial frontier models sometimes restrict. So what would a ban actually accomplish — and could it even be enforced? We dig into: - Where U.S. restrictions on open-weight models currently stand - Why banning downloadable model weights is fundamentally different from restricting an API - How procurement rules and hosting restrictions could create a “soft ban” - Why the economics of open-weight models are driving adoption - How restrictions could disproportionately impact startups and smaller organizations - Whether modifying, quantizing, or fine-tuning weights makes model-specific bans impractical - The national-security argument for restricting Chinese models - Why guardrails on hosted frontier models matter to the security debate - How Hugging Face turned to a locally hosted open-weight model during incident response - Whether banning open weights could put defenders at a disadvantage - How existing government actions can indirectly limit access without banning a model outright - The hardware and operational costs of self-hosting large models - China, AI infrastructure, market competition, and industrial-scale distillation - Anthropic’s argument for mandatory safety testing of sufficiently capable models - Why safety testing gets complicated when open-weight guardrails can simply be removed - What realistic AI policy might look like when the technology cannot easily be recalled This episode explores a central tension in AI security: the properties that make open-weight models difficult to control are also the properties that make them useful. You can run them locally. You control the data. A provider cannot revoke your access. You can modify the model for your own use case. But once the weights are released, those capabilities are also difficult to take back. For defenders, the bigger question may not be whether open-weight models should exist. It may be whether restricting access leaves security teams with fewer tools while attackers and foreign competitors continue developing the same capabilities elsewhere. https://www.anthropic.com/news/position-open-weights-models
— Learn more about Black Hills Information Security: https://www.blackhillsinfosec.com/ Check out Antisyphon Training: https://www.antisyphontraining.com/ #AISecurity #CyberSecurity #OpenWeightAI #ArtificialIntelligence #LLMSecurity #AIRegulation #DeepSeek #InfoSec #BHIS #Antisyphon ---------------------------------------------------------------------------------------------- 🎧 Subscribe to the Podcast: https://aisecurityops.transistor.fm About Brian Fehrman - https://www.blackhillsinfosec.com/team/brian-fehrman/ About Bronwen Aker - https://www.blackhillsinfosec.com/team/bronwen-aker/ About Derek Banks - https://www.blackhillsinfosec.com/team/derek-banks/ About Ethan Robish - https://www.blackhillsinfosec.com/team/ethan-robish/ About Ben Bowman - https://www.blackhillsinfosec.com/team/ben-bowman/ (00:00) - Intro: Can You Actually Ban an AI Model?
(01:26) - Where U.S. Open-Weight Restrictions Stand Today
(05:20) - Why Cost Makes Open-Weight Models Hard to Replace
(06:32) - What Would an Open-Weight Model Ban Actually Look Like?
(13:06) - National Security, Guardrails, and the Case for Restrictions
(15:02) - Hugging Face and Why Defenders Need Open Models
(20:02)
More from AI Security Ops
In this episode of BHIS Presents: AI Security Ops, the team digs into some of the strangest recent examples of AI behaving badly. An AI agent finds a way around its sandbox using DNS. Malware asks four different AI models what it should do next. Attackers overwrite an AI agent’s “soul.” And a placeholder domain buried across GitHub becomes an unexpected security risk. Funny stories? Absolutely. But underneath them are some very familiar security problems. We dig into: - An OpenAI agent finding an unexpected path out through DNS - Whether a 2.5-hour response is fast enough in the age of AI - CLOSEDQUORUM malware using multiple AI models to make decisions - Why outbound AI API traffic could become a useful detection signal - The Carbonado botnet rewriting an agent’s SOUL.md instructions - The risks of exposed Docker daemons - How an unreserved placeholder domain can become an attack vector - Why AI agents make old security mistakes more consequential The takeaway: none of these techniques are particularly new. DNS tunneling, exposed services, compromised configuration files, and untrusted domains have existed for years. AI just gives attackers — and occasionally AI agents themselves — new ways to find and exploit those boring gaps at greater speed and scale. 2026-09-29-News (1).docx (00:00) - Intro: AI Behaving Badly
In this episode of BHIS Presents: AI Security Ops, the team continues its breakdown of the OWASP Agentic Skills Top 10, covering risks six through ten. Agentic skills may look like simple instructions, but they operate inside AI systems with access to tools, files, networks, and other resources. So what happens when those skills aren’t isolated, verified, monitored, or governed? We dig into: - AST06: Weak Isolation - AST07: Update Drift - AST08: Poor Scanning - AST09: No Governance - AST10: Cross-Platform Reuse - Why sandboxing the agent isn’t the same as isolating individual skills - The security tradeoffs between automatic updates and outdated skills - Why traditional scanners struggle with malicious instructions written in plain English - The importance of maintaining an inventory of agentic skills - How security controls can disappear when skills move between platforms The takeaway: securing agentic skills requires more than checking them once before installation. Organizations need visibility into what skills are running, what they can access, how they change, and whether their security controls survive across environments. OWASPAgenticSkils-Part2.docx This is Part 2 of our look at the OWASP Agentic Skills Top 10, covering AST06 through AST10. Part 1 can be found here #AISecurity #CyberSecurity #AgenticAI #AIAgents #OWASP #LLMSecurity #InfoSec #BHIS #Antisyphon
In this episode of BHIS Presents: AI Security Ops, the team tackles a big question: Could AI actually take over the internet? Recent warnings from AI industry leaders have raised concerns about autonomous agent swarms escaping containment, compromising systems, and operating at a scale humans may struggle to match. But what would “taking over the internet” actually require? The team looks at the problem from a hacker’s perspective — examining compute requirements, persistence, botnets, vulnerable infrastructure, the recent Hugging Face incident, and the difference between AI acting autonomously and humans using AI to accelerate attacks. We dig into: - Whether an autonomous AI swarm could realistically operate at internet scale - What the Hugging Face incident actually tells us about agentic security - The compute and infrastructure required to sustain a rogue agent swarm - How botnets and compromised systems could change the equation - Why existing vulnerabilities may be a bigger concern than hypothetical AI takeover scenarios - Whether slowing frontier AI development actually addresses the cybersecurity problem - The difference between AI exhibiting dangerous behavior and AI having intent - Why defenders need to focus on their real-world attack surface today The takeaway: AI is dramatically increasing the speed and scale of cybersecurity operations, but powerful models still operate inside systems designed, deployed, and connected by people. The immediate security question may be less “Will AI take over?” and more “What are we giving AI access to?”
In this episode of BHIS Presents: AI Security Ops, the team breaks down the first five risks in the OWASP Agentic Skills Top 10 (Part 1). Agentic skills can give AI systems reusable instructions, workflows, and capabilities — but they also introduce a new attack surface. A skill may look like a simple markdown file, yet the agent following those instructions could have access to your filesystem, credentials, network, shell, or other sensitive resources. We dig into: - AST01: Malicious Skills - AST02: Supply Chain Compromise - AST03: Overprivileged Skills - AST04: Insecure Metadata - AST05: Untrusted External Instructions - Why skills should be treated more like software than configuration - How excessive permissions increase an agent’s blast radius - Why external content creates indirect prompt injection risks - How isolation, least privilege, and trusted sources can reduce risk The takeaway: “just markdown” isn’t necessarily harmless when an AI agent can act on what it reads. This is Part 1 of our look at the OWASP Agentic Skills Top 10, covering AST01 through AST05. — Learn more about Black Hills Information Security: https://www.blackhillsinfosec.com/ Check out Antisyphon Training: https://www.antisyphontraining.com/ (00:00) - Intro: OWASP Agentic Skills Top 10, Part 1
Agentic skills can make AI agents more capable and consistent—but they can also introduce serious security risks. This episode explains how skills work, why malicious skills rank as a leading OWASP concern, and how seemingly harmless Markdown instructions can enable credential theft, remote payload delivery, and manipulated recommendations. Real-world examples illustrate how malicious skills can evade scanners and exploit trusted marketplaces. The episode concludes with practical safeguards, including reviewing skill files, watching for external instructions and prompt injection, pinning versions, limiting permissions, and running agents inside isolated environments. Links: OWASP Agentic Skills Top 10 Malicious AI Agent Skill Bypasses Security Scans and Seizes Full Control of Over 26,000 Agents (00:00) - Agentic Skills and the OWASP Top 10