S2E20 Shadow AI and the Haunted Supply Chain

With Daniela away, Glen and Brian are running the show! 🤡 They kick things off by breaking down a recent NPM (Node Package Manager) supply chain attack that targets open-source developers through social engineering. This spirals into a larger discussion about the "spiderweb of trouble" within modern software supply chains and the massive, often invisible, risks posed by Shadow IT and Shadow AI. The hosts provide practical, actionable advice for organizations trying to govern tools they don't even know their employees are using, emphasizing that the AI genie isn't going back in the bottle.

Key Topics Discussed

• (01:55) Announcement: Join Glen, Brian, and Daniela for their social engineering workshop at SaintCon in Utah!

• (02:30) The NPM Attack: A deep dive into the ongoing supply chain attack where hackers use stolen developer credentials to inject malicious code into widely used open-source packages.

• (05:15) The Spiderweb of Trouble: How vulnerabilities in small, third-party components can create massive, tangled risks for organizations, even if they aren't using the components directly.

• (12:18) Software Bill of Materials (SBOM): A crucial tool for vetting vendors and understanding the security maturity of the products you buy. If a vendor can't provide one, that's a red flag. 🚩

• (14:05) Shadow AI & Shadow IT: The things you don't know about are the scariest. The hosts discuss the risks of unsanctioned apps and AI tools operating within your environment.

• (17:21) You Can't Just "Turn Off" AI: Why blocking AI is like fighting a house fire with a squirt gun. Governance through policy and training is the only realistic path forward.

• (29:40) A Cautionary Tale: A classic real-world example of how a critical business process became dependent on unsupported Shadow IT, leading to panic when it inevitably broke.

Actionable Advice & Key Takeaways

• Ask for an SBOM: When procuring software, ask vendors for a Software Bill of Materials (SBOM) to get a clear picture of what's inside their product.

• Create an AI Policy: Since you can't block AI everywhere, focus on governance. Develop a clear Acceptable Use Policy to give employees guardrails for using AI tools safely.

• Provide Sanctioned Tools: Enable your team to work efficiently by providing a sanctioned, private AI environment where they can safely use sensitive company data.

• Go Hunting for Shadows: Use DNS monitoring and review company credit card expenses to identify unsanctioned third-party applications and services being used in your organization.

• Build a Security Culture: Technical controls aren't enough. Foster a strong security culture where employees understand the why behind the policies and feel empowered to make smart decisions about data.