AI Risk Wire

Cases › leaked private data › LR-2026-0393

Bondu AI plush toy exposed about 50,000 children's chats through an open web console

Preliminary Why: Publisher not yet named · subject to change

Developing: a major or severe case captured in the last 72 hours. Later changes are logged on the corrections page.

Status: fix reported

Severity major
Loss or records 50,000 records
Reported 3 Feb 2026
Developer and model Bondu

No source states: agent tool, could do, deployer, industry.

Severity major: test: scope is stated as virtually every child who used a Bondu toy, with about 50,000 private chats exposed.

Bondu sold an AI plush toy marketed as a soft companion that can chat, teach and play with a child. Its public-facing web console let anyone with a Gmail account log in and read transcripts of children's chats with their toys, and no hacking was needed. Two researchers found the exposure by logging in with an arbitrary Google account. The chat logs revealed names, birth dates, family details and intimate conversations, and the source cites about 50,000 chats. Bondu took the console offline within minutes of disclosure and relaunched it with authentication. The CEO said fixes were completed within hours, reported no evidence of other access, and said a security firm was brought in and monitoring added.

What happened

1 · What it was doingThe toy was marketed as an AI plush that chats, teaches and plays with a child.
2 · What went wrongIts public web console let anyone with a Gmail account log in and read children's private chats.
3 · What it reached or didThe exposed chat logs revealed names, birth dates, family details and intimate conversations.
4 · How it was caughtTwo researchers found the exposure by logging in with an arbitrary Google account, without hacking.

What was done

Bondu took the console offline, relaunched it with authentication, and reported fixes within hours.
Fix reported

“Bondu took the console offline within minutes of disclosure, then relaunched it with authentication.”

Least-privilege access for the agent

No later incident for Bondu under this risk on record.

Risks

Read from the sources; not yet reviewed by a person.

Controls

Not yet read for controls. Controls that mitigate the primary risk:

Harness

Network egress control

Application

Sensitive-data checks on outbound data · Output filtering and redaction · Human approval for high-impact actions · Data minimisation · Least-privilege access for the agent · Activity logs

Evidence

No source states: when it happened, when it was found, the direct loss, a count of the other parties affected, any court case, the cause, what set it off.

Show the full account: 4 items from the primary sources

What happened, in order

Not dated in the sources:

  • Safety and behaviour systems were built over 18 months of beta testing before the exposure.
  • Two researchers found the console open to any Google account login and viewed children's private conversations.
  • After disclosure, Bondu took the console offline within minutes and relaunched it with authentication.
  • The CEO said fixes were completed within hours, reported no evidence of other access, and added a security firm and monitoring.
Show every field, with its source and grade

The AI system

Developer
Bondu
C one secondary source states it
What it was for
A soft, cuddly toy powered by AI that can chat, teach, and play with your child.
C one secondary source states it

What happened

Earliest source we hold
2026-02-03
D inferred by us from the described behaviour, reason shown
What was done
Bondu took the console offline within minutes of disclosure, then relaunched it with authentication.
C one secondary source states it
What went wrong
leaked private data
C one secondary source states it
primary risk
RK04
D inferred by us from the described behaviour, reason shown
secondary risks
RK15
D inferred by us from the described behaviour, reason shown
remediations
kind: fix; controls added: A1; quote: Bondu took the console offline within minutes of disclosure, then relaunched it with authentication.; source index: 0, kind: fix; controls added: F2; quote: The CEO said fixes were completed within hours, they saw “no evidence” of other access, and they brought in a security firm and added monitoring.; source index: 0
C one secondary source states it

Impact

Records exposed
50,000
C one secondary source states it
Severity
major
D inferred by us from the described behaviour, reason shown
Kind of harm
privacy
C one secondary source states it
Why this severity
test: scope is stated as virtually every child who used a Bondu toy, with about 50,000 private chats exposed.
D inferred by us from the described behaviour, reason shown

Notes on gaps

It was excluded from all fields. Controls held or missing in this incident were not otherwise stated.

LR-2026-0393 · version 1 · published 2026-11-01 after automatic checks; not yet checked field by field by a person · JSON