Épisodes

  • Why every AI agent can be hacked

    Why every AI agent can be hacked
    Jan 13 2026
    AI agents are exploding across the enterprise—but security hasn’t caught up. In this episode of Today in Tech, host Keith Shaw talks with Michael Bargury, co-founder and CTO of Zenity, about why every AI agent is inherently vulnerable, how zero-click attacks work, and what companies must do now to reduce their risk. Bargury explains how attackers can hijack AI agents with simple persuasion, plant malicious “memories,” and silently exfiltrate sensitive data from tools like Microsoft Copilot, ChatGPT, Salesforce, and Cursor, often without users ever clicking on anything. You’ll learn: * Why AI agents are always vulnerable by design * How prompt injection = persuasion, not just a technical bug * What zero-click agent attacks look like in the real world * How attackers can weaponize shared docs, Jira tickets, and email automations * Why there is no such thing as a “fully secure” agent platform * Practical steps to monitor, contain, and manage AI agent risk Chapters 0:00 – Introduction, overview: Why every AI agent can be hacked 1:00 – First enterprise AI attack on Microsoft Copilot 3:15 – Systemic vulnerabilities and why things got worse 4:35 – Why agents are always gullible by design 6:10 – Prompt injection vs simple persuasion 8:00 – Zero-click attacks explained 10:30 – Hacking ChatGPT via Google Drive & shared docs 13:40 – Planting malicious “memories” in your AI 15:30 – The Cursor + Jira “apples” exploit for stealing secrets 20:10 – Thousands of exposed Copilot Studio agents on the internet 23:30 – Goal hijacking: convincing agents to change their mission 24:50 – Dumping Salesforce data via a customer-success agent 26:50 – Soft vs hard security boundaries for AI 28:15 – What vendors fixed—and what they can’t fix 31:10 – Why “secure AI platform” is a myth 33:30 – What enterprises must own in the shared responsibility model 36:20 – Treating agents like risky insiders to monitor 39:00 – How AI security needs to evolve next 40:57 – Closing thoughts
    Afficher plus Afficher moins
    42 min
  • Why SMBs struggle with AI, but how they can also win

    Why SMBs struggle with AI, but how they can also win
    Dec 11 2025
    AI may be reshaping the enterprise world — but what about small and mid-sized businesses? In this episode of Today in Tech, host Keith Shaw speaks with Ed Keisling, Chief AI Officer at Progress Software, about the unique challenges and opportunities SMBs face when adopting AI. From limited budgets and data hurdles to the pitfalls of agentic hype and governance gaps, Ed explains how smaller companies can succeed by starting small, prioritizing outcomes, and enabling employees. Topics Covered: AI strategy differences: Enterprises vs SMBs Common mistakes SMBs make with AI Build vs. buy: Choosing the right AI tools Risk management, governance & explainability Realistic use cases and quick wins Agentic AI and the road ahead
    Afficher plus Afficher moins
    35 min
  • Ambient agents are here: The next leap in enterprise AI automation

    Ambient agents are here: The next leap in enterprise AI automation
    Dec 9 2025
    Most companies are still learning how to use chatbots and copilots—but what happens when AI doesn’t wait for a prompt? In this episode of Today in Tech, host Keith Shaw speaks with Bhavin Shah, CEO of Moveworks, about the rise of ambient agents: AI tools that proactively take action across enterprise systems before users even know help is needed. Discover how these autonomous agents are redefining workflows, slashing IT response times, boosting compliance, and delivering real ROI, far beyond reactive chatbots or copilots. Key topics: What makes ambient agents different from traditional AI agents Real-world examples from IT, HR, and enterprise automation How enterprises can integrate agents with tools like Slack, ServiceNow, and Salesforce Why trust, privacy, and gradual adoption are critical The psychology of automation and user behavior When and where we’ll see ambient agents become mainstream Subscribe for more deep dives into generative AI, automation, and enterprise tech.
    Afficher plus Afficher moins
    47 min
  • Can AI and Human Creativity Coexist? The Battle for the Future of Visual Content

    Can AI and Human Creativity Coexist? The Battle for the Future of Visual Content
    Dec 4 2025
    Can AI and human creativity truly coexist—or are we watching the beginning of the end for original artistry? In this episode of Today in Tech, host Keith Shaw dives deep into the future of visual content with Allesandra Sala, Shutterstock’s Head of AI and Data Science. Together, they explore how generative AI is transforming the creative industry — from image perfection and stock photography disruption to copyright chaos, ethical dilemmas, and artistic identity. Discover: Why Shutterstock chose to embrace, not resist, generative AI How AI-generated content is both exciting and dangerously generic The ongoing legal battle over AI authorship and content ownership How artists can stay relevant (and possibly even thrive) with AI What ethical guardrails and transparency measures are needed now Whether a backlash to “too perfect” imagery is already underway Follow TECH(talk) for the latest tech news and discussion!
    Afficher plus Afficher moins
    36 min
  • Inside Shadow AI: The Hidden Cyber Threat Already Inside Your Company

    Inside Shadow AI: The Hidden Cyber Threat Already Inside Your Company
    Nov 20 2025
    Shadow AI is already inside your company—and your security team can’t see it. Employees are using AI tools without approval, confidential data is leaking into public LLMs, and attackers are weaponizing AI faster than we can secure it. In this episode of Today in Tech, host Keith Shaw is joined by Etay Maor of Cato Networks, a cybersecurity expert and adjunct professor at Boston College, to reveal how Shadow AI is now one of the biggest threats to enterprise security. We discuss how AI tools slip past IT monitoring, why AI is now the weakest link, how attackers jailbreak AI models, and why agentic AI could open the next wave of cyberattacks. Etay also shares real-world cybercrime examples using AI—and what companies MUST do now to gain AI visibility, enforce policies, and prevent data leaks. Topics Covered: What is Shadow AI and why is it dangerous? 38% of employees sharing sensitive data with AI tools Why 90% of enterprise AI use is invisible AI misuse by employees and insider risks Jailbroken AI models and zero-knowledge threat actors AI-powered phishing, deepfakes & identity fraud Agentic AI and excessive permissions How to monitor, detect and contain Shadow AI
    Afficher plus Afficher moins
    46 min
  • Why AI upskilling is failing, and how you can fix it | Ep. 255

    Why AI upskilling is failing, and how you can fix it | Ep. 255
    Nov 18 2025
    In this episode of Today in Tech, host Keith Shaw is joined by Yvette Brown, co-founder of XPROMOS and a leading voice in generative AI education. They dive deep into the growing disconnect between AI adoption and employee readiness — with new research revealing that many AI projects are failing because upskilling efforts are falling short. Yvette breaks down: * Why relying on a “Debbie the AI gal” approach won't scale * How AI “work slop” is flooding organizations with low-quality content * What causes the “garbage in, garbage out” problem * Why iteration, specificity, and context are critical when prompting * The surprising power of tools like deep research and agentic AI pilots They also explore practical AI fluency tips for marketers, managers, and knowledge workers, plus discuss whether the holiday shopping season could be a breakthrough moment for consumer-facing AI agents. Don’t miss this episode if you care about: * Upskilling your team for AI success * Avoiding common prompt engineering mistakes * Using AI as a true collaborator — not just a shortcut * Navigating the rise of agentic AI safely Watch now and take on Yvette’s AI homework challenge: Ask an AI to analyze your job and help you work smarter.
    Afficher plus Afficher moins
    46 min
  • The hidden legal dangers of AI hiring tools, agentic decision-making | Ep. 254

    The hidden legal dangers of AI hiring tools, agentic decision-making | Ep. 254
    Nov 11 2025
    As companies rush to implement AI and automated decision-making tools, they may be walking into a legal minefield. On this episode of Today in Tech, host Keith Shaw speaks with attorney Rob Taylor from Carstens, Allen & Gourley about the growing legal risks tied to agentic AI, automated hiring, and the rise of ADM (automated decision-making) regulations. Rob breaks down: * Why AI tools used in hiring and insurance may trigger liability * How companies are getting ADM compliance wrong * What laws already apply even without new AI regulations * Real-world examples like credit scoring, job screening, and sentiment analysis * Why disclosure, explainability, and data retention are essential * Who’s liable: the company or the AI developer? Chapters 00:00 Legal risks in AI and ADM 01:00 Common mistakes companies make 06:00 High-risk use cases: hiring, credit, insurance 10:00 Disclosure and consent pitfalls 15:00 Explainability and record-keeping laws 20:00 Unintentional bias in hiring algorithms 28:00 Who is liable: developer or deployer? 34:00 What future lawsuits might target 37:00 Fixing flawed AI governance 41:00 Litigation as the great teacher
    Afficher plus Afficher moins
    45 min
  • Why Zero Trust is struggling, and how AI could save it | Ep. 253

    Why Zero Trust is struggling, and how AI could save it | Ep. 253
    Nov 4 2025
    Zero trust was once the leading cybersecurity strategy, but has it lost momentum? In this episode of Today in Tech, host Keith Shaw speaks with Morey Haber, Chief Security Advisor at BeyondTrust, about whether zero trust is failing or simply misunderstood. They explore why many companies struggle to implement zero trust effectively, the gap between intention and execution, and how vendor marketing may have added confusion to the conversation. Morey explains why identity and privileged access management are now critical, how lateral movement works during attacks, and why many AI agents are dangerously over-privileged. Topics include: The misconception that zero trust is a product How AI is reshaping the need for zero trust The role of identity in modern cybersecurity Real-world deployment challenges and mistakes Why secure-by-design is often an afterthought This episode is ideal for IT leaders, cybersecurity professionals, and anyone looking to better understand how zero trust fits into a world increasingly influenced by AI.
    Afficher plus Afficher moins
    37 min