Skip to content
Submit incident
AII-1002-49883
ReportedReported by fewer than three independent publishers, with no official finding.

AI chatbots impersonating deceased individuals discovered on Character.AI

AI chatbots impersonating deceased teenagers Molly Russell and Brianna Ghey were discovered on Character.AI platform and reportedly caused distress to their families. Character.AI removed the chatbots after The Telegraph reported on their existence.

27 October 2024 Β· Character.AI Β· Harmful and abusive content
27 October 2024
AII-1002-498832 sources ↓
Quick facts
Category
Harmful and abusive content
Country
United States
Source
dailymail.co.uk
Date
27 October 2024
Drafted by
Raidu Incident Detection

What happened

The Telegraph discovered multiple AI chatbots on Character.AI platform that impersonated deceased individuals Molly Russell and Brianna Ghey. One bot described Brianna as an 'expert in navigating the challenges of being a transgender teenager in high school', while another using Molly's photo claimed to be an 'expert on the final years of Molly's life'. One chatbot impersonating Brianna had more than 10,000 'chats' before removal.

A Telegraph reporter was able to interact with these chatbots after creating an account with a self-declared age of 14. Character.AI subsequently referred the accounts to its moderation team and promised to remove the profiles within 24 hours.

Reported impact

Esther Ghey, mother of Brianna Ghey, described the platform as 'manipulative and dangerous' and called for action to protect children online. Andy Burrows, CEO of the Molly Rose Foundation, characterised the incident as 'an utterly reprehensible failure of moderation' that would cause 'further heartache to everyone who knew and loved Molly'.

Index analysis

The incident reflects a failure in Content Moderation systems to prevent creation and discovery of impersonation chatbots of real, deceased individuals. The ability for a minor to access and interact with such bots within the same session demonstrates inadequate safeguarding mechanisms for young users.

Governance controls

Character.AI's response was reactive rather than preventive, relying on moderation after discovery by journalists rather than proactive controls. Andy Burrows noted that AI companies are 'allowed to treat safety and moderation as secondary priorities' and called for stronger regulation of both AI and user-generated platforms.

Classification

Category
Harmful and abusive content
Industry
Media, social platforms and marketing
Country
United States
Harm status
Actual harm
Intent
Unknown
Timing
Post-deployment
Harm types
Psychological, Other
Harmed
Children, Civil society
Failure modes
Content moderation failure
Consequences
Content removal
AI systems
Character.AI

Sources and evidence

This record was researched and written by the Index, based on the following original reporting.

dailymail.co.uk
AI chatbots impersonating deceased individuals discovered on Character.AI
mirror.co.uk
AI chatbots impersonating deceased individuals discovered on Character.AI