Event language
UI language
**In 2025, WMF merged its security and anti-abuse product teams into what is now the Product Safety and Integrity team, or PSI.** **Since then, our team has overhauled account security, and deployed new systems for detecting bots and abusive activity – while taking the time along the way to make the wikis easier to use. Now we’re building on that progress to introduce new protections, while integrating AI more deeply into our security work.** **We’ll start this talk by discussing what’s changed on the wikis over the last year, and our strategy for making them easy on users and getting them shipped. This includes:** * **Making two-factor authentication open to all, and requiring it for sensitive roles…** * **…which can now be done quickly and passwordlessly with passkeys.** * **Replacing our weak, annoying CAPTCHA with mostly-invisible bot detection…** * **…that provides signals that get incorporated into tools for checkusers and stewards to spot abusive activity.** **Then we’ll talk briefly about what we’re up to next: How do we protect users and their accounts at scale when even the code is customizable? How can we use the dramatic advancements that have been made in AI-assisted security analysis to our advantage?** **Finally, we’ll hold 5-10 minutes for Q&A.**
You need to log in before contacting the organizer.
