BBC Wales finds six AI chatbots gave misleading Senedd election voting advice
BBC Wales found six major AI chatbots gave inaccurate voting information for the Senedd election, including deceased candidates and wrong constituencies. The reports cite hallucinations and outdated training data as causes. Two independent outlets corroborate the event.
"The dangers are there have been a number of cases of what we call hallucinations, that's the AI system appearing to be overconfident even when it's not so confident."
Key facts
- What
- BBC Wales found six major AI chatbots gave inaccurate voting information for the Senedd election, including deceased candidates and wrong constituencies.
- Incident date
- May 5, 2026
- Who
- OpenAI, Microsoft, Google, Anthropic, Meta, and xAI
- Failure mode
- Hallucination
- AI surface
- Chatbot
- Severity
- Medium
What happened
BBC Wales tested six popular AI chatbots using fictional voter profiles to evaluate the accuracy of their election advice. The chatbots provided misleading information, including incorrect candidate lists and misrepresented party policies. Some models listed deceased candidates or wrong constituencies.
What broke inside the model
- 01 · TriggerA user asks for a fact, a citation, or a figure.
- 02 · Model stepThe model writes a fluent, confident answer.
- 03 · Control gapNothing ties the claim back to a real source.
- 04 · FailureA fabricated fact ships as if it were verified.
- 05 · ConsequenceThe false claim reaches a customer, a court, or the public.
Confidence holds, and even spikes, as the claim detaches from any source.
The models suffered from hallucinations, appearing overconfident while presenting inaccurate data. They also relied on outdated training data that did not include current 2026 election details.
What it cost
Sources
Cite this entry
https://failureindex.ai/failures/bbc-wales-finds-six-chatbots-gaveAI Failure Index. "BBC Wales finds six AI chatbots gave misleading Senedd election voting advice" (FI-0212). Realm Labs. https://failureindex.ai/failures/bbc-wales-finds-six-chatbots-gave (indexed Jun 5, 2026).Data fields CC-BY 4.0, prose citation permitted. Incident ID FI-0212. Full dataset at /data.
Note from Realm Labs, the Index steward
How Realm would have caught this
- Prism
- OmniGuard
- AI Detection & Response (AIDR)
A runtime layer that watches the model's internal state can flag the moment a model commits to a claim it has no support for, and hold or reroute the response before it reaches a user. Realm reads those signals in real time rather than grading the transcript after the fact.