BU BÖLÜM HAKKINDA
How do professionals exercise judgment when their eyes andears can be deceived? In this episode of "Decoding Ethics," host Jeanne Viljoen speaks with Christian Scott, a "White Hat" ethical hacker and founder of Tantalum Security, whose team tests firms' defenses against deepfake attacks. The focus is preparation rather than fear: what professional accountants, firms and business leaders need to understand abouthow deepfakes are changing fraud and cyber risk.
Christian traces how social engineering has moved fromphishing, phone calls and text messages to AI-enabled deepfakes. Live deepfakes were still blocky and blurry in 2022, deepfakes made in advance are now nearlyimpossible to detect. He describes how, in authorized assessments, his team clones faces and voices to test wire transfer controls. He also walks through deal-room fraud, in which attackers swap wire instructions and route fundsthrough bank accounts opened in someone else's name by defeating KYC checks. Other schemes he covers include deepfake extortion, market manipulation and fake kidnapping calls that use children's voices taken from social media.
On defenses, Christian explains why preventing these attacksgoes far beyond multi-factor authentication, from watching for signals such as impossible travel to using identity proofing tools firms may already have. He warns that liveness checks on video calls may not hold up for long, and that verification will have to rely on knowledge, not just appearance. Above all, he argues for a culture with no exceptions to the process, backed by layered controls rather than reliance on individual employees.
The conversation closes on governance. Christian calls fortransparency about where AI is used in an organization and for building governance into processes by design. For smaller firms, he recommends starting with incident response and business continuity. He leaves listeners with a question: What does trust mean for your organization, and can it be audited?
🔗 More on Christian and Tantalum Security: https://tantalum.security/
📊 Deepfake statistics: University at Buffalo, January 2026: https://bit.ly/3UtK1iX
🛡️ Cybersecurity Awareness Month 2026:
- NCA, "Don't Make It Easy for Them": https://bit.ly/3WslKu9
- European Cybersecurity Month, ENISA, "Think Before U Click": https://bit.ly/4hW7Fwd
⚖️ EU AI Act, European Commission: https://bit.ly/3VCNlsu
Note: The EU AI Act was adopted in 2024. Its transparency rules apply from Aug. 2, 2026.
🖼️ Swiss cheese model illustration: Ben Aveling, via Wikimedia Commons, licensed under CC BY-SA 4.0(https://creativecommons.org/licenses/by-sa/4.0).Adapted from the original. Model developed by James Reason. https://bit.ly/4joQzK0
📝 Correction: At 15:20, the hacker in the 2024 KnowBe4 incident is described as Chinese state-backed. KnowBe4 identified the individual as a North Korean fake IT worker: https://www.knowbe4.com/press/knowbe4-issues-warning-to-organizations-after-hiring-fake-north-korean-employee
Follow Decoding Ethics for more conversations on trust,technology, and ethics – and share it with colleagues.#DecodingEthics #AccountingEthics #Cybersecurity#CybersecurityAwarenessMonthChapter Titles
00:00 Introduction to Christian Scott and the Topic of AIDeception
01:52 What a White Hat Ethical Hacker Actually Does
04:51 Capabilities and Evolution of Deepfake Technology
06:26 Most Convincing AI Impersonations and SecurityAssessments
09:16 Real-World Example of Deepfake Fraud in FinancialServices
14:00 Challenges to Traditional KYC and Remote Onboarding
16:48 Other Uses of AI Deception Beyond Wire Fraud
19:34 Ethical Responsibilities and Accountability in AIDeception
23:26 The Importance of Transparency and Governance in AIUse
25:49 The Role of Organizational Culture and EmployeeTraining
27:55 Good Governance and Risk Management in AI Adoption
30:20 Practical Steps for Small Organizations to ImproveCybersecurity
32:20 The Future Outlook on AI and Trust
