Whisper Leak: How Threat Actors Can See What You Talk to AI About

To kick off Season 3 of Microsoft Threat Intelligence Podcast, host⁠ ⁠⁠Sherrod DeGrippo is joined by Microsoft security researchers Anna Seitz and Jonathan Checchi.   Our guests examine two developments shaping today’s threat landscape: the cloud-native evolution of ransomware group Storm-0501 and the SesameOp backdoor’s abuse of trusted AI platforms for stealthy command-and-control. The discussion highlights how identity, hybrid-cloud pivot points, and federated authentication enable high-impact attacks without traditional malware, and why policy-compliant platform abuse is becoming harder to detect.   Sherrod, Anna, and Jonathan provide guidance for defenders around enforcing MFA, tightening conditional access and identity controls, monitoring across cloud and on-prem environments, and partnering with platform providers to disrupt emerging attacker tradecraft.  In this episode you’ll learn:       What happens when threat actors gain control of highly privileged identities  Why monitoring identity behavior is as critical as monitoring endpoints  How attacker tactics are adapting to environments that blend cloud and on-prem systems   Some questions we ask:      What does recent threat activity tell us about where the landscape is headed?  How is Storm-0501 using federated authentication in their operations?  What should security teams focus on as AI becomes more integrated into systems?  Resources:   View Anna Seitz on LinkedIn   View Sherrod DeGrippo on LinkedIn    Related Microsoft Podcasts:                    Afternoon Cyber Tea with Ann Johnson  The BlueHat Podcast  Uncovering Hidden Risks        Discover and follow other Microsoft podcasts at microsoft.com/podcasts     Get the latest threat intelligence insights and guidance at Microsoft Security Insider    The Microsoft Threat Intelligence Podcast is produced by Microsoft, Hangar Studios and distributed as part of N2K media network.

In this episode of the Microsoft Threat Intelligence Podcast, host⁠ ⁠⁠Sherrod DeGrippo is joined by security researchers Geoff McDonald and JBO to discuss Whisper Leak, new research showing that encrypted AI traffic can still unintentionally reveal what a user is asking about through patterns in packet size and timing.   They explain how LLM token streaming enables this kind of side-channel attack, why even well-encrypted conversations can be classified for sensitive topics, and what this means for privacy, national-level surveillance risks, and secure product design. The conversation also walks through how the study was conducted, what patterns emerged across different AI models, and the steps developers should take to mitigate these risks.  In this episode you’ll learn:       Why packet sizes and timing patterns reveal more information than most users realize  How user-experience choices like showing streamed text create a larger attack surface  The difference between classic timing attacks and the new risks uncovered in Whisper Leak    Resources:   View JBO on LinkedIn  View Geoff McDonald on LinkedIn    View Sherrod DeGrippo on LinkedIn    Learn more about Whisper Leak     Related Microsoft Podcasts:                    Afternoon Cyber Tea with Ann Johnson  The BlueHat Podcast  Uncovering Hidden Risks        Discover and follow other Microsoft podcasts at microsoft.com/podcasts     Get the latest threat intelligence insights and guidance at Microsoft Security Insider      The Microsoft Threat Intelligence Podcast is produced by Microsoft, Hangar Studios and distributed as part of N2K media network.

In this episode of the Microsoft Threat Intelligence Podcast, host⁠ ⁠⁠Sherrod DeGrippo is joined by Matt Duncan, Vice President of Security Operations and Intelligence at the North American Electric Reliability Corporation’s E-ISAC, to explore the cyber threats targeting the North American power grid. Matt breaks down why the grid remains resilient despite increasing pressure from nation-states, cybercriminals, and hacktivists, how AI is lowering the barrier of entry for attackers, and why OT systems and interconnected devices present unique risks.   He also highlights real success stories, the value of large-scale grid exercises, and how strong collaboration and a focus on foundational security practices help defenders keep power flowing safely and reliably.    In this episode you’ll learn:       How severe weather events trigger heightened cyber-readiness across utilities  What motivates hacktivist groups and how their tactics differ from other threat actors  Why outdated equipment and legacy systems remain such attractive targets  Some questions we ask:      Are you seeing more educated and capable OT-focused adversaries now?  How do you work with policymakers to help them understand these threats?  If you could eliminate one misconception about securing the grid, what would it be?    Resources:   View Matt Duncan on LinkedIn   View Sherrod DeGrippo on LinkedIn   Learn more about E-ISAC    Related Microsoft Podcasts:                    Afternoon Cyber Tea with Ann Johnson  The BlueHat Podcast  Uncovering Hidden Risks        Discover and follow other Microsoft podcasts at microsoft.com/podcasts     Get the latest threat intelligence insights and guidance at Microsoft Security Insider    The Microsoft Threat Intelligence Podcast is produced by Microsoft, Hangar Studios and distributed as part of N2K media network.

In this episode of the Microsoft Threat Intelligence Podcast, host⁠ ⁠⁠Sherrod DeGrippo is joined by security researchers Tori Murphy and Anna Seitz to unpack two financially motivated cyber threats. First, they explore the Payroll Pirates campaign (Storm 2657), which targets university payroll systems through phishing and MFA theft to reroute direct deposits. Then, they examine Vanilla Tempest, a ransomware group abusing fraudulent Microsoft Teams installers and SEO poisoning to deliver the Oyster Backdoor and Recita ransomware.   Together, they discuss how attackers exploit trust in identity, code signing, and SaaS platforms and share practical steps organizations can take to strengthen defenses, from phishing-resistant MFA to stricter executable controls and out-of-band banking verification.  In this episode you’ll learn:       How Payroll Pirates diverted university salaries through SaaS HR phishing schemes  Why universities are prime targets for identity-based cyberattacks  How Vanilla Tempest evolved from basic ransomware to complex multi-stage attacks  Some questions we ask:      How are attackers stealing credentials and paychecks?  Why do attackers create inbox rules after compromising accounts?  What alerts should organizations monitor for these types of attacks?  Resources:   View Tori Murphy on LinkedIn   View Anna Seitz on LinkedIn  View Sherrod DeGrippo on LinkedIn   Investigating targeted “payroll pirate” attacks affecting US universities  Microsoft Threat Intelligence healthcare ransomware report highlights need for collective industry action    Related Microsoft Podcasts:                    Afternoon Cyber Tea with Ann Johnson  The BlueHat Podcast  Uncovering Hidden Risks        Discover and follow other Microsoft podcasts at microsoft.com/podcasts     Get the latest threat intelligence insights and guidance at Microsoft Security Insider    The Microsoft Threat Intelligence Podcast is produced by Microsoft and distributed as part of N2K media network.

In this episode of the Microsoft Threat Intelligence Podcast, host⁠ ⁠⁠Sherrod DeGrippo is joined by Zack Korman, CTO of cybersecurity startup Pistachio. They explore the reality of AI in security, cutting through hype to discuss where AI is both brilliant and flawed, how vendors AI-wash outdated tech, and why Zack believes AI won’t replace jobs but instead scale human creativity. They also dive into phishing simulations, human psychology behind social engineering, AI-powered attacks, jailbreak chaining between AI systems, and the future risks and opportunities AI introduces in cybersecurity.   In this episode you’ll learn:       How to evaluate whether a vendor is truly using AI in their product  The psychology behind why people fall for phishing attacks  Why human judgment will remain essential in the era of AI-driven security.  Some questions we ask:      How can AI unlock new capabilities in cybersecurity?  What questions should people ask AI security vendors?  Why do trained security professionals still fall for phishing attacks?  Resources:   View Zack Korman on LinkedIn   View Sherrod DeGrippo on LinkedIn     Related Microsoft Podcasts:                    Afternoon Cyber Tea with Ann Johnson  The BlueHat Podcast  Uncovering Hidden Risks        Discover and follow other Microsoft podcasts at microsoft.com/podcasts     Get the latest threat intelligence insights and guidance at Microsoft Security Insider    The Microsoft Threat Intelligence Podcast is produced by Microsoft and distributed as part of N2K media network.

In this episode of the Microsoft Threat Intelligence Podcast, host Sherrod DeGrippo is joined by Chloé Messdaghi and Crane Hassold to unpack the key findings of the 2025 Microsoft Digital Defense Report; a comprehensive look at how the cyber threat landscape is accelerating through AI, automation, and industrialized criminal networks.  They explore how nation-state operations and cybercrime have fused into a continuous cycle of attack and adaptation, with actors sharing tooling, infrastructure, and even business models. The conversation also examines AI’s growing impact, from deepfakes and influence operations to the defensive promise of AI-powered detection, and how identity compromise has become the front door to most intrusions, accounting for over 99% of observed attacks.  Listeners will gain perspective on:  How AI is shaping both attacker tradecraft and defensive response.  Why identity remains the cornerstone of global cyber risk.  What Microsoft’s telemetry—spanning 600 million daily attacks—reveals about emerging threats and evolving defender strategies.  Questions explored:  How are threat actors using AI to scale deception and influence operations?  What does industrialized cybercrime mean for organizations trying to defend at scale?  How can defenders harness AI responsibly without overreliance or exposure?    Resources:   Download the report and executive summary  Register for Microsoft Ignite  View Chloé Messdaghi on LinkedIn  View Crane Hassold on LinkedIn  View Sherrod DeGrippo on LinkedIn   Related Microsoft Podcasts:                    Afternoon Cyber Tea with Ann Johnson  The BlueHat Podcast  Uncovering Hidden Risks        Discover and follow other Microsoft podcasts at microsoft.com/podcasts     Get the latest threat intelligence insights and guidance at Microsoft Security Insider    The Microsoft Threat Intelligence Podcast is produced by Microsoft and distributed as part of N2K media network.

In this episode of the Microsoft Threat Intelligence Podcast, host⁠ ⁠⁠Sherrod DeGrippo is joined by Tori Murphy, Anna Seitz, and Chuong Dong to break down two threats: the modular backdoor PipeMagic and Medusa ransomware. They discuss how PipeMagic disguises itself as a ChatGPT desktop app to deliver malware, its sophisticated modular design, and what defenders can do to detect it.   The team also explores Medusa’s evolution into a ransomware-as-a-service model, its use of double extortion tactics, and the broader threat landscape shaped by ransomware groups, social engineering, and the abuse of legitimate tools.   In this episode you’ll learn:       Why modular malware is harder to detect and defend against  How attackers abuse vulnerable drivers to disable security tools  Why leak sites play a central role in ransomware operations  Some questions we ask:      How did Microsoft researchers uncover PipeMagic in the wild?  Why do ransomware groups often borrow names and themes from mythology?  What initial access techniques are commonly associated with Medusa attacks?    Resources:   View Anna Seitz on LinkedIn  View Chuong Dong on LinkedIn    View Sherrod DeGrippo on LinkedIn     Related Microsoft Podcasts:                    Afternoon Cyber Tea with Ann Johnson  The BlueHat Podcast  Uncovering Hidden Risks        Discover and follow other Microsoft podcasts at microsoft.com/podcasts     Get the latest threat intelligence insights and guidance at Microsoft Security Insider    The Microsoft Threat Intelligence Podcast is produced by Microsoft and distributed as part of N2K media network.

In this episode of the Microsoft Threat Intelligence Podcast, host⁠ ⁠⁠Sherrod DeGrippo is joined by Kelly Bissell, Corporate Vice President at Microsoft, to explore how domain impersonation and typosquatting are changing in the age of AI.  They discuss how attackers are increasingly using AI and bots to scale online deception, why this tactic is so effective, and how Microsoft is countering cutting-edge defenses like Siamese neural networks to detect fraudulent domains in real time. Kelly shares insights on the massive scale of these threats, the shift toward defender advantage, and the broader implications for securing organizations worldwide.  In this episode you’ll learn:       How attackers use AI and bots to scale domain impersonation and typosquatting  Why defenders may finally have the higher ground in the fight against online fraud  How Microsoft’s Siamese neural network model detects fraudulent domains in real time  Some questions we ask:      What excites you most about this new detection approach?  How do fake domains fit into a larger social engineering chain?  What indicators should defenders watch for in typosquatting domains?    Resources:   View Kelly Bissell on LinkedIn   View Sherrod DeGrippo on LinkedIn     Related Microsoft Podcasts:                    Afternoon Cyber Tea with Ann Johnson  The BlueHat Podcast  Uncovering Hidden Risks        Discover and follow other Microsoft podcasts at microsoft.com/podcasts     Get the latest threat intelligence insights and guidance at Microsoft Security Insider

In this episode of the Microsoft Threat Intelligence Podcast, host⁠ ⁠⁠Sherrod DeGrippo is joined by Microsoft researchers Kelsey Clapp and Anna Seitz to examine two major cybercrime campaigns. The team unpacks Storm 2561’s use of SEO poisoning to distribute Trojanized software like SilentRoute and Bumblebee, stealing VPN credentials and paving the way for ransomware brokers.  They also dive into Storm 1811’s ReadBed malware, a loader deployed through bold social engineering tactics, such as fake IT help desk calls via Teams, that enable lateral movement and ransomware deployment. The discussion highlights how modern threat actors exploit trust, extend attack chains, and continually evolve their techniques, underscoring the importance of vigilance, strong security controls, and verifying before trusting. In this episode you’ll learn:      How Storm 2561 uses SEO poisoning to trick users into downloading Trojanized software The role of trust, urgency, and habit in social engineering tactics Practical steps organizations can take to block these threats and strengthen defenses Some questions we ask:     Why are initial access loaders such a big risk for organizations? How are threat actors using fake IT help desk calls to gain access? What steps should defenders take to cut off these entry points? Resources:  View Anna Seitz on LinkedIn View Kelsey Clapp on LinkedIn  View Sherrod DeGrippo on LinkedIn  Related Microsoft Podcasts: Afternoon Cyber Tea with Ann Johnson The BlueHat Podcast Uncovering Hidden Risks     Discover and follow other Microsoft podcasts at microsoft.com/podcasts  Get the latest threat intelligence insights and guidance at Microsoft Security Insider The Microsoft Threat Intelligence Podcast is produced by Microsoft and distributed as part of N2K media network.

In this episode of the Microsoft Threat Intelligence Podcast, host⁠ ⁠⁠Sherrod DeGrippo is live from Black Hat 2025 with a special lineup of Microsoft security leaders and researchers. First, Sherrod sits down with Tom Gallagher, VP of Engineering and head of the Microsoft Security Response Center (MSRC). Tom shares how his team works with researchers worldwide, why responsible disclosure matters, and how programs like Zero Day Quest (ZDQ) are shaping the future of vulnerability research in cloud and AI security. He also announced the next iteration of ZTQ with $5 million up for grabs. Next, Sherrod is joined by Eric Baller (Senior Security Researcher) and Eric Olson (Principal Security Researcher) to unpack the fast-changing ransomware landscape. From dwell time collapsing from weeks to minutes, to the growing role of access brokers, they explore how attackers operate as organized ecosystems and how defenders can respond. Finally, Sherrod welcomes Travis Schack (Principal Security Researcher) alongside Eric Olson to examine the mechanics of social engineering. They discuss how attackers exploit urgency, trust, and human curiosity, why AI is supercharging phishing campaigns, and how defenders can fight back with both training and technology. In this episode you’ll learn:     How MSRC partners with researchers across 59 countries to protect customers Why Zero Day Quest is accelerating vulnerability discovery in cloud and AI How ransomware dwell times have shrunk from days to under an hour Resources: View Sherrod DeGrippo on LinkedIn  Zero Day Quest — Microsoft Microsoft Security Response Center Blog Related Microsoft Podcasts:   Afternoon Cyber Tea with Ann Johnson The BlueHat Podcast Uncovering Hidden Risks      Discover and follow other Microsoft podcasts at microsoft.com/podcasts Get the latest threat intelligence insights and guidance at Microsoft Security Insider   The Microsoft Threat Intelligence Podcast is produced by Microsoft and distributed as part of N2K media network.