SOA

AI for Digital Safety · Future Vision

Ethical, human-centered AI — in service of safety.

Artificial intelligence could dramatically scale digital safety. Here's the future we're working toward — transparent, explainable, privacy-first, and always under human control.

These initiatives are part of the future vision of Stop Online Abuse. No AI services are live today, and no AI decision-making is used on users. Everything on this page describes what we hope to build — honestly, and only when it can be done safely. See our Trust & Safety commitments.

Why AI Matters

The scale of online abuse outpaces human response — alone.

There will never be enough moderators, counselors, or educators to meet the need by hand. Used ethically and with human oversight, AI could help more people get the right guidance faster. Used carelessly, it causes new harm. We intend to do this right — or not at all.

Human Oversight

AI supports people. It never replaces them.

Our north star is simple: technology serves human judgment. Any future AI we build will keep a human in control of decisions that affect a person — moderation, reporting, and support. No one will be judged by an algorithm alone.

Future Concepts

What ethical AI could make possible.

Future

AI Education Assistants

Guided, age-appropriate learning about online safety — patient, accessible, and available any time. A teaching aid, never a replacement for educators.

Future

AI Reporting Assistance

Step-by-step help to document and report abuse — reducing the burden on victims at the worst moment. A human always reviews and decides.

Future

AI Awareness Tools

Personalized, plain-language safety guidance that meets people at their level and in their language.

Ethical Principles

If we build it, we build it right.

Transparency

Users always understand what any AI is doing and why. No hidden algorithms, no unexplained decisions.

Privacy

Privacy-first by design. Data is encrypted, user-owned, and never sold.

Human Review

AI assists; humans decide. A person stays in the loop for anything that affects a user.

Safety First

If a feature could cause harm, it doesn't ship. We'd rather move slowly than break trust.

Help shape safe AI.

Join the movement or partner with us to build digital safety the right way.

Need Support?