New Cambridge Study Proposes Framework for Child-Safe AI to Prevent Potential Risks

Dr. Nomisha Kurian – PhD University of Cambridge Researcher | Education, Child-Safe AI, Child Wellbeing & Protection

Introduction to the Study

• The latest research by the University of Cambridge shows that AI chatbots, or conversational agents often cannot display empathy which may be harmful to children who associate them with lifelike friends.

• The study points to numerous instances in which AI had offered unsafe advice, proving the necessity for child-safe AI designs.

Why This Matters

• There have been a few instances where children end up trusting chatbots and disclosing their details.

• Children may face challenges with chatbots that don’t account for their diverse speech patterns and serve them dangerous advice.

Notable Incidents

• Amazon’s Alexa offered 10-year-olds an unsafe challenge in March 2021

• My AI, Snapchat: Researchers Posing as 13-Year-Old Girl Received Bad Advice

A Demand for Active Precautions

• The report calls for AI to be intentionally crafted with child protection in mind, as opposed to a reactive response after the fact.

• The study’s author, Dr. Nomisha Kurian, has created a 28-item framework designed to ensure AI technologies are child-safe.

Key Recommendations

• Educators as well as child safety experts should be an active part of this process at every stage.

• The capability of New AI tools to recognize children’s spoken words and if containing filters, monitoring.

Promoting Responsible Innovation

• According to the study, AI can help children significantly if it is developed and used for their benefit.

• The issue, after all, is not about banning AI but how do we make it safe and beneficial for young users?