Sign up to save your podcasts
Or
Share S2 E4: Words as Weapons - The Psychology and Exploitation of Talking to AI
Share to email
Share to Facebook
Share to X
Share to email
Share to Facebook
Share to X
Or
S2 E4: Words as Weapons - The Psychology and Exploitation of Talking to AI
AI is changing how we work and how we get hacked.
Every day, employees rely on large language models to draft emails, summarize data, and solve problems. But the same conversational power that makes them useful also makes them vulnerable. The words we use can shape, persuade, or even exploit the systems we trust.
In this episode, we speak with Elle Neal, AI and data scientist at Databutton and associate at AI4C, and Ibai Castells, Senior Hive Member at CovertSwarm, about how communication, psychology, and offensive security collide in the world of AI.
Together, they explore:
- How human tone and phrasing can cause AI performance to spiral or improve
- Why polite and hostile prompts reveal both emotional intelligence and risk
- How attackers are turning everyday language into prompt injection exploits
- What “AI as a new employee” means for culture, trust, and continuous testing
This isn’t about fear. It’s about awareness.
As AI becomes part of every team, the way we talk to it could define how secure or exposed we really are.
Successful companies don’t just adopt AI.
They learn to challenge it continuously.
They subscribe to CovertSwarm.
--
Successful companies are constant targets for attackers. Those who take security seriously don’t test their defenses once a year.
They subscribe to CovertSwarm.
View all episodes
By CovertSwarmAI is changing how we work and how we get hacked.
Every day, employees rely on large language models to draft emails, summarize data, and solve problems. But the same conversational power that makes them useful also makes them vulnerable. The words we use can shape, persuade, or even exploit the systems we trust.
In this episode, we speak with Elle Neal, AI and data scientist at Databutton and associate at AI4C, and Ibai Castells, Senior Hive Member at CovertSwarm, about how communication, psychology, and offensive security collide in the world of AI.
Together, they explore:
- How human tone and phrasing can cause AI performance to spiral or improve
- Why polite and hostile prompts reveal both emotional intelligence and risk
- How attackers are turning everyday language into prompt injection exploits
- What “AI as a new employee” means for culture, trust, and continuous testing
This isn’t about fear. It’s about awareness.
As AI becomes part of every team, the way we talk to it could define how secure or exposed we really are.
Successful companies don’t just adopt AI.
They learn to challenge it continuously.
They subscribe to CovertSwarm.
--
Successful companies are constant targets for attackers. Those who take security seriously don’t test their defenses once a year.
They subscribe to CovertSwarm.