In this final episode of the Events Demystified seven-year anniversary miniseries, host Anca Platon Trifan explains that “shadow AI” is often ordinary employees using tools like ChatGPT, Copilot, Gemini, Claude, or browser extensions to work faster, unintentionally creating data paths outside company controls.

She argues shadow AI reflects both employee responsibility and leadership gaps, citing Verizon’s 2026 DBIR data on rising AI use, unauthorized services, and common exposed data types like source code. The episode covers AI security best practices: clear data classification with real examples (especially for events), minimum-necessary data sharing, risks from extensions and excessive agent permissions, the need for limited autonomy, auditability, a kill switch, vendor and integration due diligence, transparent monitoring, and safe incident reporting. Trifan shares a Bill.com virtual card fraud experience and closes with questions leaders should ask and a podcast break until September.

👉 AI hallucinations (July 24), and

👉 Shadow AI, security best practices (July 31).

Produced by Tree-Fan Events Production LLC

Book Anca for speaking or consulting:⁠⁠⁠⁠⁠⁠https://ancaplatontrifan.me/booking-page/⁠⁠⁠⁠⁠⁠

Follow for more on AI, events, and leadership:

LinkedIn: ⁠⁠⁠⁠⁠⁠https://www.linkedin.com/in/ancatrifan/⁠⁠⁠⁠⁠⁠

Instagram: ⁠⁠⁠⁠⁠⁠https://www.instagram.com/anca.platon.trifan/⁠⁠⁠⁠⁠⁠

Original podcast music by Fable Score Music

Tree-Fan Events is a woman-owned boutique event production agency focused on delivering high-impact experiences through the FIT4EVENTS™ framework.

In this seven-year anniversary episode of the Events Demystified Podcast, Anca Platon Trifan uses live event production examples to show how confident, premature conclusions can waste time, then connects that dynamic to AI hallucinations: confident but false “confabulations” such as fabricated citations, links, dates, and quotes. She explains how large language models generate plausible text without guaranteeing truth, highlights the 2023 Mata v. Avianca case where lawyers filed non-existent cases generated by ChatGPT, and argues hallucinations reveal existing gaps in human critical thinking. The episode covers why polished answers feel credible (illusory truth effect, confirmation bias), why critical thinking matters more than prompt engineering, and how overreliance scales risk in event work (contracts, permits, schedules, accessibility, bios, stats). She outlines verification practices: assess consequence, identify key claims, return to original sources, avoid “AI-to-AI” verification, preserve uncertainty, and limit agent authority based on impact.

👉 AI hallucinations (July 24), and

👉 Shadow AI, security best practices (July 31).

Produced by Tree-Fan Events Production LLC

Book Anca for speaking or consulting:⁠⁠⁠⁠⁠https://ancaplatontrifan.me/booking-page/⁠⁠⁠⁠⁠

Follow for more on AI, events, and leadership:

LinkedIn: ⁠⁠⁠⁠⁠https://www.linkedin.com/in/ancatrifan/⁠⁠⁠⁠⁠

Instagram: ⁠⁠⁠⁠⁠https://www.instagram.com/anca.platon.trifan/⁠⁠⁠⁠⁠

Original podcast music by Fable Score Music

Tree-Fan Events is a woman-owned boutique event production agency focused on delivering high-impact experiences through the FIT4EVENTS™ framework.

Host Anca Platon Trifan marks the Events Demystified Podcast’s seven-year anniversary miniseries by continuing last week’s discussion on AI bias into a broader AI ethics focus on accountability. Using her own AI agent that scans her Notion speaking database, finds calls for speakers, drafts pitches, and posts updates to Slack/Discord, she explains how risk changes when AI moves from recommending to acting, especially representing someone without review. She argues AI can do work but cannot absorb responsibility, and warns organizations often scatter accountability across leaders, vendors, and teams. Ethical AI becomes real in concrete permissions, incentives, auditability, transparency with recourse, and meaningful human oversight. She highlights data permission/ownership, workforce impacts, shadow AI, and the need for a named leader who owns outcomes, sometimes choosing not to use AI.

The five-Friday July mini-series on AI will cover:

👉 AI bias and stereotypes (July 10),

👉 AI ethics and accountability (July 17),

👉 AI hallucinations (July 24), and

👉 Shadow AI, security best practices (July 31).

Produced by Tree-Fan Events Production LLC

Book Anca for speaking or consulting:⁠⁠⁠⁠https://ancaplatontrifan.me/booking-page/⁠⁠⁠⁠

Follow for more on AI, events, and leadership:

LinkedIn: ⁠⁠⁠⁠https://www.linkedin.com/in/ancatrifan/⁠⁠⁠⁠

Instagram: ⁠⁠⁠⁠https://www.instagram.com/anca.platon.trifan/⁠⁠⁠⁠

Original podcast music by Fable Score Music

Tree-Fan Events is a woman-owned boutique event production agency focused on delivering high-impact experiences through the FIT4EVENTS™ framework.