Designing Out Harm: Trust and Safety Guide

Free guide for social, gaming and AI agent platforms: the harms your product enables, the UK OSA, EU DSA and AI Act duties landing now, and how to design safety in.

A platform's guide to designing in trust and safety

A field guide for social platforms, gaming companies and creators of AI agents on the harms their products enable, the regulators now watching, and the operating model that can keep them out of the firing line.

Share this guide

Send it to Trust & Safety, Legal or Product leaders who need it.

FREE PDF GUIDE

Delivered instantly to your browser once you complete the form.

Download the guide

Complete the form to receive the PDF instantly. Work email required.

EU-hosted CRM. No third-party ad tracking on this form.

Why this book, why now

Regulators are handing out landmark fines. Courts are ruling against major platforms. AI has made new harms possible overnight. Designing Out Harm helps you get ahead of it before you become the next statistic.

€120m

EU fine against X, December 2025

The first non-compliance decision issued under the Digital Services Act.

$6m

Jury verdict against Meta and Google, March 2026

The first Social Media Adolescent Addiction bellwether trial to reach a verdict.

380%

Rise in AI-generated CSAM reports in a single year

Reported by the Internet Watch Foundation.

$4.4tn

Estimated annual cost of financial crime

Larger than the entire economy of the United Kingdom.

Who this is for

Written for the people who carry the consequences when a product decision turns into an incident, an investigation or a headline.

Trust & Safety and Risk leaders

Who need to know exactly where their product is exposed, and which harms their own design decisions make possible.

Legal and Compliance teams

Navigating four overlapping regulatory regimes at once: the UK Online Safety Act, the EU DSA and AI Act, Australia's codes, and US state and federal rules.

Product and engineering leaders

Building social, gaming or AI agent platforms who need to design harm out from the start rather than bolt safety on afterwards.

Inside the guide

How ordinary product features become harm vectors, from open DMs to friend suggestions to agent memory.
A map of the four regulatory regimes now converging on platforms, and what each one actually asks you to evidence.
The new harms AI agents introduce, including synthetic media, persuasive companionship and automated fraud.
An operating model for trust and safety that scales with the product instead of trailing behind it.
The metrics, records and reviewer trails regulators expect to see when they ask you to prove your controls work.

Tuteliq exists to make online spaces safer by design

We give platforms the ability to detect behavioural risk before it becomes harm, surfacing the patterns that signal grooming, bullying, fraud, radicalisation and exploitation, so the platforms people rely on every day can intervene early rather than clean up after the fact.

Tuteliq is a single behavioural risk detection API, built to ship UK Online Safety Act, EU DSA, EU AI Act, Australian and US KOSA-aligned compliance out of the box. Our models are developed alongside criminologists on real exploitation patterns, processing millions of interactions daily across 13 harm categories, one detection endpoint each, covering grooming, cyberbullying, self-harm, fraud and more, at 1.1-1.5s typical response times with zero content retention.

Questions about the guide

What is Designing Out Harm?

Designing Out Harm is a field guide for social platforms, gaming companies and creators of AI agents. It sets out the harms their products enable, the regulators now watching, and an operating model for building trust and safety into the product rather than adding it afterwards.

Who should read this guide?

It is written for Trust & Safety, Product, Legal and Risk leaders. It assumes you own decisions about product design, compliance posture or incident response, and it focuses on what those roles need to act on rather than general awareness material.

Which regulations does the guide cover?

The UK Online Safety Act, the EU Digital Services Act, the EU AI Act, Australia's online safety codes, and US rules including KOSA-aligned obligations and state-level statutes. The guide explains where those regimes overlap and where their requirements diverge.

Does it cost anything?

No. The guide is free. You complete a short form with your work email and the PDF downloads immediately.

What happens to the details I submit?

They are stored in our EU-hosted CRM and used to send you the guide and related updates. You can unsubscribe from any email we send, and our Privacy Policy sets out the full basis for processing.

Get your copy of Designing Out Harm

Free, no sales call required. The PDF downloads as soon as the form is submitted.