SecurePod: Making AI and cyber safety feel human, not homework

The eighth episode of SecurePod features Anna Pieczatkowska, Information Security Training and Awareness Manager at Sonova and founder of Cyber Aperitivo, where she tests AI tools to make cyber safety practical rather than intimidating. In conversation with host Nick Kelly, she speaks about what actually works when raising security awareness inside a large organization. Anna traces her habit of testing AI tools most people only meet through TikTok influencers, and makes the case that basic account hygiene, like enabling MFA, gets skipped precisely because nobody's teaching it. They explore how to keep awareness programs engaging without turning them into theater, why LLM hallucinations are a real risk once the topic turns sensitive, and how tricks like "emoji smuggling" show prompt injection is a live, evolving threat. The conversation turns to the gap between a vendor stating it follows security standards and an organization actually being able to prove it, and why AI governance has to pair user training with vendor accountability. Anna closes by unpacking her work with ENISA's AR-in-a-Box initiative, a two-year effort to help organizations build and sustain security awareness programs at scale, and offers a practical habit: use an LLM to interrogate a tool's terms and conditions instead of trusting an influencer's word for it.