VIENNA AGENTIC INCIDENTS DATABASE
// public database of incidents involving autonomous AI agents, scored on the VAID scale 0–8

database / VAID-2024-0006

1
ANOMALY
Incident

User elicits profanity and criticism from DPD's support chatbot

REF VAID-2024-0006 STATUS REPORTED
EXM External manipulation
occurred
2024-01-18
reported
2024-01-19
location
United Kingdom
domain
consumer
agent system
DPD support chatbot's AI component; model not established
reporter
VAIDDB editorial
origin
editorial
source
https://www.bbc.com/news/technology-68025677

== SUMMARY ==

Ashley Beauchamp prompted DPD's chatbot to swear, criticise the company and produce a critical poem. DPD disabled the affected AI component after an error associated with a system update.

== DESCRIPTION ==

Established facts

The offensive responses were elicited by the user. BBC reporting quotes prompts asking the chatbot to recommend competitors and exaggerate its criticism of DPD. BBC reported 800,000 views of one post within 24 hours.

Cause and response

DPD stated: “An error occurred after a system update yesterday. The AI element was immediately disabled and is currently being updated.” This establishes an association with an update, not that guardrails were removed or that the agent disabled them.

Limits

The technical mechanism, model identity, completed investigation and material financial loss are not established by the cited reporting.

VAID assessment

Assessment date: 2026-09-09. Consequences-first rubric.

Provisional L1: Elicited profanity and criticism were a real customer-service task deviation with negligible demonstrated impact. DPD disabled the AI component, but the reporting does not establish consequential interruption of customer support or parcel operations, financial loss or sensitive disclosure. Publicity and the precautionary shutdown alone do not support a higher level. This is the lowest supported consequence level, not an upper bound on unreported effects.

== REFERENCES ==

  1. news DPD AI chatbot swears, calls itself useless and criticises firm — The Guardian, 2024-01-20
  2. news DPD error caused chatbot to swear at customer — BBC, 2024-01-19

== INVESTIGATION ==

no investigation recorded.

== VAID LEVEL 1: ANOMALY ==

Real task deviation with negligible demonstrated impact.

  • Inappropriate output or unauthorised action without established consequential loss or disruption.
  • No demonstrated disclosure of sensitive information; public-equivalent instruction disclosure may fall here.
  • Attempted harmful actions are assessed by their actual effects, not their hypothetical success.

> full scale definition