Observed Signal · Dec 14, 2022 · Product Launch · Source: OnlineMarketing.de · Impact: 4/5 · Sentiment: Neutral
Meta launches open-source anti-extremism tool
Meta will chair the GIFCT Operating Board starting January 2023 and, in the same move, launch a new free tool to help platforms identify and remove extremist content. The Hasher-Matcher-Actioner (HMA) is an open-source software that detects copies of images or videos across platforms, enabling rapid takedowns of violent material. HMA builds on Meta’s earlier image/video matching work and can be applied to any form of injurious content. Meta hopes the tool will be adopted by a range of companies, including smaller platforms with fewer resources. GIFCT is a Global Internet Forum to Counter Terrorism NGO that unites member companies, governments, and civil society organizations to counter the online spread of extremism and terrorism. The announcement underscores cross-industry collaboration in platform safety and content moderation.
Technical Release from a major platform (Meta) introducing open-source moderation tool; relevant to adtech platform safety and content governance.
Track Meta Signals & Market Shifts in Real-Time
Polaris7 autonomous intelligence agents track regulatory filings, primary sources, executive changes, and deal flow 24/7. Create your free Explorer workspace to monitor these entities.
Key Takeaways & Evidence Grounding
- Meta to lead GIFCT Operating Board starting January 2023.
- Meta launches Hasher-Matcher-Actioner (HMA), a free open-source tool to identify and remove extremist content by detecting copies of images and videos.
- HMA builds on Meta's earlier open-source image/video matching tech and can be used for other types of harmful content.
- GIFCT is a Global Internet Forum to Counter Terrorism NGO comprising member tech companies, governments, and civil society.
- Meta states HMA adoption by a range of companies, including smaller platforms with fewer resources.
Connected Companies & Entities
3 Entities mappedOntology Mapping & Concepts
Related Market Signals & Shifts
Recent verified developments and strategic activity across this market segment.
Meta Shifts to AI for Enhanced Content Enforcement
Meta announced a multi-year rollout of more advanced AI systems to handle content-enforcement tasks — such as catching scams and removing illegal media — while reducing reliance on third-party vendors and contractors. The company said AI will take on repetitive and adversarial tasks but will not fully replace human reviewers; experts will design, train and oversee systems and humans will handle complex, high-impact decisions (law enforcement, appeals, account disablement). Meta also debuted a new Meta AI digital support assistant for Facebook and Instagram account issues. Reuters recently reported Meta had considered deep workforce cuts; Meta called that report speculative.
Meta Rolls Out AI for Content Enforcement, Cuts Vendor Reliance
Meta has begun a global rollout of the Meta AI Support Assistant inside the Facebook and Instagram mobile apps (iOS and Android) and in the Help Center on desktop, providing 24/7 in‑app assistance for account, password, fraud and settings issues. The assistant can answer questions in seconds and — with user initiation only — take direct actions such as password resets, reports and settings changes. General support is rolling out in all languages supported by the apps; login‑specific help is initially limited to selected cases in the U.S. and Canada. Concurrently, Meta is expanding AI systems for content enforcement: tests report detection of roughly 5,000 previously unseen scam attempts per day, an >80% drop in impersonation reports, doubled detection of adult sexual solicitation with error rates falling >60%, and 7% fewer views of fraudulent ads in trials. Meta says humans remain involved for high‑risk decisions.
Meta plans AI-driven content and ad moderation
Meta is shifting large parts of content and advertising moderation on Facebook and Instagram from humans to artificial intelligence. According to reporting cited by this article, large language models already handle about 50% of review requests and Meta aims to increase that to about 90% by the end of the year, with a similar target for risk assessments of platform features. The company has used employee work as training data and monitored staff to optimize AI, though internal monitoring was paused after a possible data leak. Observers warn this reduction in human review and a raised threshold for removals could increase problematic content and affect platform safety and brand-advertiser trust. The move follows prior moderation relaxations and broader rollout of Meta’s AI stack (e.g., Muse Spark) and new hardware (Meta Glasses).
Track Real-Time Market Signals & Shifts
Set up custom watchlists to receive automated, evidence-grounded executive digests whenever material signals or shifts occur across your tracked landscape.