Meta & Children's Online SafetyCIPGO Identity Integrity Pillar 2021-H1 Roadmap
Raw text
Case 4:22-md-03047-YGR Document 3029-21 Filed 05/08/26 Page 1 of 3 Exhibit 141 Case 4:22-md-03047-YGR Document 3029-21 Filed 05/08/26 Page 2 of 3 CIPGO Identity Integrity Pillar 2021-H1 Roadmap TL;DR of plans for H1 2021 In H1 2021 we will: 1. Increase agility by consolidating our Private Impersonation and High Profile Impersonation detection and response systems into a reusable system (including atomizing FB PI measurement), migrating our Page Publisher Authorization (PPA) product to a platformized architecture, and set the groundwork to migrate all our manual ID review capacity to an autonomous pool (IDV as a Service). 2. Build efficiency and decrease our ID verification backlog from ~500k to <6k jobs through automation (both auto-reject and auto-accept). 3. Hold the line on PI/HPI prevalence while mitigating impersonation risk ahead of SOAP launch. Increase coverage of high risk inauthenticity products (Broad Distribution initially) from 0.3% to 3% of global VPVs. 4. Hold the line on ID verification E2E accuracy at 80% and <3% prevalence of fake IDs in automated decisions. Make progress in fully migrating our accuracy measurement to GTM, reaching Operational status. 5. Continue to drive Legitimacy of our systems by: a. Prove out the value of authenticity where identity checks are deployed to achieve distinct goals, measuring the impact of authenticity on hypothesised harms/benefits. b. Increase our investment on Underage, making sure that our reporting systems meet regulatory expectations as well as gain credibility. c. Progress on privacy-respecting alternative verification methods including Selfie standalone (to be shipped in Novi, IG Access and Dating), Open ID Connect and NFC. Priority trade-offs (expand to all problems at the pillar level): C D High priority Medium priority B Prevalence A B 1 Legend: Top priority A 1 Problem Impersonati on 2 High Profile High Pri: Hold the line on current surfaces but invest in mitigations for new surfaces (cross-app, E Low priority C Accuracy Med pri: Validate GTM metrics and expand coverage to 80% D Cost/Efficiency Top Pri: 50% of time towards consolidating ou systems into a single re- useable system which ca CONFIDENTIAL META3047MDL-065-00258000 Case 4:22-md-03047-YGR Document 3029-21 Filed 05/08/26 5 DE 6 6 Other: design, 10 content, business development, TPM, PPM, Product Growth, SME 7 Total: 92 95 106 Page 3 of 3 ID review. This includes Underage appeals Review of the Account ~780k account photos to only determine if user is <13 years old, and if Open Questions / Topics for Discussion Question 1/ There were a few questions about the flows and backlogs for IDV and Underage. Below is comprehensive view of these flows. They are owned by different teams and mitigated in different ways. TO RATIFY: Are you happy with our goals? Description Review Type All jobs that are only ID only Reason for Size Goal Mitigation backlog ~200k <6k late inventory IDR War Room Review jobs by 7days/ 48hours [HYPERLINK Retrospective on Wednesday, we'll report back Minimum "https://www.inte investment in u13 malfb.com/tasks in previous so, user is check pointed June User aiming to change their age within and across Account only TBC; TBC need to /?t=76089081" [TBC] by \h] TBC; investigating quarters Minimum investment in u13 be inconsistency of in previous the 18 year old investigat ID use in quarters threshold (not ed protocols including U13) Various review flows ID + ~51k No goal No mitigations at Account this stage Question 2/ Sounds like the Underage Taskforce may potentially drive big changes to the H1 roadmap, do we know will we drop in case it becomes a significant investment in H1? There are 2 types of requests coming from the Age Verification Task Force: a. Incremental product changes- these have been incorporated into our roadmaps for H1 already. Besides product work, there will be increased ops needs here. Examples of these changes include adding in app 3rd CONFIDENTIAL ΜΕΤΑ3047MDL-065-00258026