Fair moderation is hard but fair, scalable moderation is harder
Throughout my career, I’ve struggled with the problem of how to enforce policies in a fair, accurate, and scalable way. A new research paper reminds us just how difficult that is
A reader asks: What should be on my ‘red line’ list?
Most T&S professionals—whether they admit it or not—have a line they won’t cross for their company. But when you're in the middle of a major, public failure, it can be hard to know what to do. Here’s my take on what to consider before quitting.
What I heard at the T&S Summit in London
My first time attending a big T&S event outside of the US brought with it a lot of fun. But I left without as deep an understanding about British or European attitudes to online safety as I'd have liked
Making the best of a T&S incident (part two)
A Trust & Safety crisis doesn’t just risk reputational damage — it grabs leadership’s attention too. That’s your moment to make the case for investment in your team. Here’s how I would turn a T&S incident into a strategic win.
Are T&S professionals part of the problem?
A new report argues that, without industry-wide standards or codes of practice, T&S professionals are vulnerable to corporate pressures and destined to always be reactive to company's conflicting priorities. The answer? Greater independence.
Some personal news, as they say
Much of the work I've done over the years has circled around the problem of how to balance human moderation and automation. Now I'll be doing it within an AI moderation company
How to get buy-in for T&S (part one)
Companies tend to invest in Trust & Safety when there's a brand or reputation risk, which isn't a structured or sustainable way to build a safety programme. Using a framework, like the Hierarchy of Growth Business Needs, can help to get buy-in before there's an emergency
How to write better transparency reports
Mandated transparency reports are a cornerstone of online safety regulation. But dozens of companies still produce voluntary disclosures that take significant time and energy to pull together. What is the purpose of these reports now that so much other information is in the public domain?
Is it time to unite T&S and AI ethics?
There's enormous potential in bringing together these often siloed disciplines and organisational functions. Given the complex, intertwined risks of AI and human interaction, it may even be a necessity.
My five Trust & Safety mantras for difficult moments
The moment we find ourself in feels politically and socially fraught meaning that T&S workers need to look out for themselves more than ever. Here's what I say to myself during times of stress
