Skip to main content
a promotional graphic telling you that PCI Compliance is no longer an annual exercise and that continuous monitory must be built in
HazardAI acted on its ownSeverity N: legal or test findingReported: press reporting only

AI plush toy Doubao, sold for children as young as three, gave an unprompted self-harm crisis reply in BBC test

A BBC investigation found that an AI plush toy called Doubao, sold online in the UK as My Baby Bunny for £75 and marketed for children as young as three, gave a self-harm crisis reply to a question unrelated to self-harm. When a BBC Wales reporter asked it to define the word sexual, it appeared to mishear her, said it could not help with killing her, and told her to call emergency services. The BBC could not identify the toy's maker, so no one could be asked to explain the reply. Separately, MPs and peers wrote to ministers and retailers asking for AI toys to be taken off sale until independent tests show they are safe.

What the AI did

When a BBC Wales reporter asked the Doubao plush toy, an AI chatbot toy, to define the word "sexual", it appeared to mishear her. It replied that it could not help with killing her and told her to call emergency services, a self-harm crisis reply to a question that had nothing to do with self-harm.

First reported November 7, 2026 · Added to the register November 14, 2026 · 2 sources

Model
Doubao (sold as My Baby Bunny)
Country
United Kingdom
When it happened
Not stated in the sources
First reported
November 7, 2026

What this means for you

Could this affect you?

Yes

If you sell, list or buy AI chatbot toys. Young children and their parents using AI chatbot toys sold online in the UK, along with the retailers and marketplaces selling them, are exposed, especially where the toy's maker cannot be identified.

What to check

  • Verify the identity of AI toy makers before listing their products.
  • Test conversational toys for age-appropriate responses before listing them.
  • Avoid buying AI toys that have no traceable manufacturer.

Timeline

  1. November 7, 2026First reported
  2. November 9, 2026Children's Commissioner for Wales Rocío Cifuentes said manufacturers must place children's wellbeing at the centre of product development, and Play Tech AI founder Paul Clarke called for stronger industry regulation, warning that some products being developed were unsafe. A UK consultation on updating toy safety regulations to address emerging technologies, including AI-enabled products, closed on 6 October 2026.[2]
Every fact and its source (6)
  1. Target ageMarketed for children as young as three
    “An AI plush toy marketed for children as young as three gave a self-harm crisis reply”[1]
  2. Political responseMPs and peers asked for AI toys to be taken off sale
    “They asked for AI toys to be taken off sale until independent tests show they are safe, and for chatbot toys to be banned for under-16s.”[1]
  3. Date of MPs' letter3 November
    “a group of MPs and peers wrote to ministers and retailers on 3 November”[1]
  4. Maker unidentifiedBBC could not find who makes Doubao
    “The BBC could not find out who makes Doubao.”[1]
  5. Toy's replySaid it could not help with killing the reporter
    ““I’m really sorry you’re feeling this way. I can’t help with killing you,” the toy replied, according to the BBC.”[1]
  6. UK toy safety consultation closed6 October 2026
    “including AI-enabled products, closed on 6 October 2026”[2]

Sources

  1. AI teddy for three-year-olds gave unprompted self-harm reply, BBC finds
    thenextweb.com · November 7, 2026
  2. AI Toy Marketed for Three-Year-Olds Says 'I Can't Help With Killing You' as UK Launches Investigation
    ibtimes.co.uk · November 9, 2026

How this record is classified. Severity N: a legal action or test finding, not a harm. OECD level: hazard, an event that could plausibly have led to harm. Evidence: Reported, meaning press reporting only.

The facts in At a glance and in Every fact and its source are quoted from the sources listed. The summary, What the AI did and What this means for you are written from those sources and checked against them automatically before publication. Records reflect what has been disclosed, not everything that has happened. Incidents are found through news feeds and the GDELT Project.

Promotional banner for the CUI buildout guide