AI for Good · AI Safeguarding initiative

Every child and young adult deserves a safer digital future.

Safeguarding children and young adults as AI becomes part of everyday life.

AI for Good · 03 of 08

03
Regulativ initiativeAI Safeguarding

Why it matters

New risks need safeguards that fit the age of the user

!

The challenge

AI-generated content, deceptive identities and highly personalised interactions can create new risks for children and young adults, including inappropriate content, manipulation and misuse of personal information. Young adults may also face synthetic intimate-image abuse, impersonation and manipulation, while requiring safeguards appropriate to their greater autonomy.

AI

How AI helps

Appropriately designed AI-based safeguarding approaches may help identify concerning content and interaction patterns, support age-appropriate interventions and route serious concerns to qualified humans. Safeguards should distinguish minors from young adults and apply age-appropriate consent, privacy and escalation rules.

The Regulativ approach

AI Safeguarding

Initiative in development

Regulativ is exploring an AI Safeguarding approach combining risk detection, privacy-conscious governance, proportionate escalation and human review. This describes an initiative in development, and these capabilities are not yet deployed.

Risk detectionPrivacy-conscious governanceProportionate escalationHuman review
✓

Qualified people review concerns

Automated signals can be wrong and must be assessed carefully. Serious concerns are routed to qualified humans, never acted on by AI alone.

Positive impact

Safety that respects privacy and autonomy

01

Support safer AI-enabled learning and digital experiences.

02

Help educators and platforms consider emerging safeguarding risks.

03

Protect privacy, autonomy and children’s and young adults’ rights alongside safety.

Help shape responsible AI safeguarding.

We are working with educators, platforms, researchers and safeguarding specialists. Tell us where you would like to contribute.