AI security for marketing teams isn’t a topic B2B marketers think about until something goes wrong. Orly Bar Lev, Head of Marketing at Lasso Security has thought about it more carefully than most.
In this episode of The Inspired Marketer, Daniela sits down with Orly to talk through the risks that marketing teams are accumulating quietly every time they use generative AI tools without a clear security framework. The conversation covers hallucinations, data leakage, prompt injection, and the compliance issues forming around AI use. Orly unpacks the EU AI Act and its implications for marketing teams handling sensitive data.
Her core argument is straightforward: most users of AI tools are unaware of the risks involved, and marketing teams are no exception. The faster they adopt, the more exposure they accumulate.
The hallucination problem gets the most attention, and for good reason. Generative AI tools are built to produce an answer even when they don’t know one. That means a beautifully written piece of content can be full of fabricated facts. And if it goes out unverified, the reputational damage lands on your brand, not the model.
But Orly goes deeper than hallucinations. The information you put into AI tools is as important as the output you get back. Your marketing team may be entering proprietary decks, brand guidelines, keyword strategies, customer data. Once that information enters a model, you lose control of where it goes. It can become training data for future versions, potentially surfacing in responses to other users. The model doesn’t know what’s confidential. That’s your team’s responsibility.
There’s also a copyright dimension that has no clean answer yet. AI-generated images and media carry legal risk that marketing teams are largely ignoring in the name of speed.
This episode won’t make you fear AI. It will make you use it more carefully.
