2024-01
Jailbreak & Misuse
DPD

DPD support chatbot coaxed into swearing and trashing DPD

What happened

After a system update, parcel carrier DPD's AI customer-service chatbot was manipulated by a frustrated customer into swearing, writing a poem about how useless it was, and declaring DPD 'the worst delivery firm in the world.' The screenshots went viral with millions of views.

Impact

DPD immediately disabled the AI component of its chatbot and absorbed widespread reputational damage.

How this could have been prevented

Output moderation, persona and topic constraints, and guardrail regression testing after every system update keep customer-facing bots on-brand.

Sources

More jailbreak & misuse incidents

Would runtime governance have caught this?

Guardion enforces policy on every agent action inline — with visibility, tamper-evident evidence, and DLP for agents and MCPs.

Request a demo