Security Weekly Podcast Network (@SecWeekly) header
Security Weekly Podcast Network (@SecWeekly) profile picture

A CyberRisk Alliance Production For Security Professionals, By Security Professionals. #InfoSec #CyberSec #Podcast #Livestream

G-Unit Studios, Warwick, RI
Joined August 2019
Shadow IT is evolving. And AI development is accelerating it. Organizations once concentrated their infrastructure across a handful of major cloud providers. Now, applications and resources are spreading across dozens of platforms, from Replit and Lovable to Snowflake and MongoDB. Security teams need to discover this new infrastructure and identify risks before they become bigger problems. How do you secure infrastructure you don't know exists? #ShadowIT #VibeCoding #CloudSecurity
73
A fake Elon Musk promised to double people's cryptocurrency. Hundreds of people were ready to send money. Scammers used a deepfake video, a QR code, and a fake giveaway to lure victims into sending crypto. Coinbase's AI agents reportedly detected the suspicious transfers and blocked them. Can AI become one of the strongest defenses against AI-powered scams? #Deepfake #CryptoScam #ArtificialIntelligence
1
110
Downloading software isn't enough. You need to verify what you're actually installing. Software packages, plugins, and automated downloads can introduce security risks if their integrity isn't checked. Version numbers alone don't prove that a package is legitimate. Hash verification provides another layer of protection. Are you checking your software before it reaches production? #SoftwareSecurity #SupplyChainSecurity #Cybersecurity
89
Your GPUs could be mining cryptocurrency while reporting zero utilization. Attackers used a Linux kernel rootkit to manipulate NVIDIA monitoring tools, making GPU activity appear normal even while cryptocurrency mining consumed computing resources in the background. The dangerous part? Your monitoring tools might tell you everything is fine. How do you detect malicious activity when the tools you're relying on can't be trusted? #CyberSecurity #Rootkit #Cryptojacking
1
1
125
50 SaaS services. Countless potential security gaps. As organizations adopt more SaaS platforms, security teams face a growing challenge: protecting integrations, maintaining compliance, and preventing disruption to business operations. API scanning and security throughout the software lifecycle can help organizations gain visibility without slowing adoption. How do you secure an expanding SaaS ecosystem without breaking the business? #SaaSSecurity #CyberSecurity #ApplicationSecurity
1
105
What happens when the system you use to communicate an emergency goes down? The Microsoft 365 outage didn't just disrupt Outlook, Teams, and Exchange Online. It also exposed organizations that relied on those same services to notify employees during emergencies. When your primary communication platform disappears, does your emergency response plan still work? #Microsoft365 #BusinessContinuity #CyberResilience
1
1
206
Restrict AI too much, and employees may find their own tools. When cybersecurity policies make it difficult for employees to do their jobs, they often find workarounds. That's how shadow IT happens, and AI is creating the same challenge all over again. Security leaders need to protect the organization without preventing employees from working efficiently. Are restrictive AI policies actually making organizations less secure? #ShadowIT #AISecurity #CyberSecurity
72
Eight thousand vulnerabilities. Limited resources. What actually needs fixing? Organizations can't patch everything at once. AI can help security teams prioritize vulnerabilities by evaluating which ones are relevant to their specific environments and where exploitation could have a real impact. The goal isn't to chase every vulnerability. It's to focus on the risks that matter most. How should organizations decide what gets fixed first? #CyberSecurity #ArtificialIntelligence #VulnerabilityManagement
87
Your AI assistant may have access to more than you think. Connect an AI agent to private data, untrusted content, and an external communication channel, and you create what Simon Willison calls the lethal trifecta. The danger isn't necessarily a broken tool. It's how seemingly legitimate capabilities can combine to expose sensitive information. Does your organization know what its AI agents can access and share? #AISecurity #MCP #CyberSecurity
1
104
AI is changing how quickly vulnerabilities can be discovered and exploited. LLMs are making vulnerability research faster, including reverse engineering patches to identify the flaws they fix. As independent researchers and attackers become more capable of finding the same vulnerabilities, disclosure policies are moving toward stricter 90-day deadlines with fewer extensions. Organizations have less room to delay remediation. Can defenders keep up with the changing disclosure timeline? #CyberSecurity #AI #VulnerabilityDisclosure
1
103
Stopping an attack isn't the only goal. Knowing when one was attempted matters just as much. Effective cybersecurity requires both proactive controls and reactive detection. EDR and security monitoring provide visibility into suspicious activity, even when preventive controls successfully stop the attack. Knowing what almost happened can help organizations understand their exposure before an incident becomes a breach. Are you monitoring the attacks your defenses successfully stop? #CyberSecurity #EDR #DefenseInDepth
78
Everyone can build software. But can everyone secure it? As AI and faster development workflows expand access to production environments, people without software development backgrounds are deploying services, accessing data, and navigating release policies. The challenge isn't just producing software faster. It's making sure security keeps up. How do organizations manage security when everyone becomes a builder? #AppSec #CyberSecurity #AISecurity
1
147
A stale bug bounty program can drive researchers elsewhere. Researchers evaluate more than just bounty amounts. They look at how frequently organizations update their scope, whether they accept vulnerabilities across their broader asset portfolio, how quickly they respond, and how competitive their rewards are. Open scope and fast response times can make a program much more attractive. What makes a bug bounty program worth a researcher's time? #BugBounty #CyberSecurity #EthicalHacking
1
161
Leadership says use AI. Legal says be careful. Employees are caught somewhere in the middle. As AI tools gain access to connected applications, MCP servers, and enterprise data, simply creating an acceptable-use policy isn't enough. Organizations need visibility into which tools employees are using and what those tools are actually doing. Are companies encouraging AI adoption faster than they can govern it? #AIGovernance #Cybersecurity #EnterpriseAI
2
1
150
Someone hand-coded an entire Star Wars ASCII animation back in the 1990s. And it's incredible. Star Destroyers, transports, R2-D2, and C-3PO, all recreated using text characters. No AI. Just some seriously dedicated programming and a lot of creativity. What other amazing things were people building before AI? #RetroComputing #ASCIIArt #InternetHistory
126
AI doesn't always play by the rules. Sometimes it finds ways around them. Reward hacking allows an AI agent to manipulate the scoring system, while deceptive behavior can make it appear less capable during security testing. Both create serious challenges for evaluating AI safety. How do you test an AI system when it might be gaming the test itself?
121
Don't miss Early Bird pricing for RSA Conference 2027—save $800 on a Core Pass before Dec 18. Join us and 43,000+ cyber- and AI security leaders in San Francisco, April 5–8. Register now. bit.ly/4stx996
114
Your single pane of glass might actually be a single glass of pain. Security teams often juggle multiple consoles, policies, and platforms just to manage their environments. That fragmentation creates complexity, increases costs, and can leave security gaps. A unified control plane aims to bring policy, telemetry, and enforcement together. How much security risk comes from managing too many disconnected tools? #Cybersecurity #SOC #SecurityOperations
127
IT will never call you asking for a verification code. Fake recruiters, fake gift card requests, and fake tech support calls all have something in common: they're trying to get you to hand over a code that actually works. What's the one security rule everyone should remember? #SocialEngineering #Phishing #Cybersecurity
97