About the Global 100 Forum
The Global 100 Forum is a question and answer community for people who have to make real decisions about AI-generated content. Teachers deciding what to do with a flagged essay. Editors checking whether a submitted photo is what it claims to be. Trust and safety teams sizing up a new detection vendor. Students who were accused of using a chatbot and want to know what the evidence actually shows.
It sits alongside the Global 100 Index, which ranks detection and content authentication platforms. The forum is where the practical questions that a ranking cannot answer get worked through in public, with sources.
What we cover
There are four sections, and each one has a narrow job.
- AI text detection. How detectors score writing, what their error rates look like in independent tests, and how to read a percentage without treating it as proof.
- Deepfakes and synthetic media. Synthetic faces, cloned voices and manipulated video. What people can spot by eye, what they cannot, and where forensic tools help.
- Academic integrity. Detector flags in coursework, fair process for students, and policy choices for instructors and departments.
- Provenance and C2PA. Content Credentials, watermarking and signed capture. The shift from guessing whether something is synthetic to proving where it came from.
Who answers
Early threads are written by forum staff. Every staff post carries a visible Staff label next to the author name, so you always know when you are reading the forum team rather than another member. As members join, their answers appear without that label, and the best answer in each thread is picked on its merits, not on who wrote it.
How we handle sources
A claim about accuracy, bias or detection performance should come with something you can check. Staff answers link to the study, specification or institutional statement they rely on, and we prefer primary sources: the paper rather than the press release, the standard rather than a blog summary. If a source is weak, preliminary or a vendor's own marketing, the answer says so.
Numbers age quickly in this field. Detectors are retrained, generators improve, and a test from two years ago may not describe today's tools. Answers carry dates for that reason, and we update threads when a newer source changes the picture instead of quietly editing old claims.
Corrections and disagreement
If you think an answer is wrong, say so in the thread and show your source. Disagreement backed by evidence is the most useful thing a forum like this can produce. Staff will correct their own posts when they are wrong and note the change in the thread rather than rewriting history.
What this forum is not
It is not a place to get a verdict on a specific person, a way to accuse someone, or a help desk for any one product. We do not run detectors on documents people paste in, and we remove posts that try to identify or shame private individuals.
Contact
Questions about the forum itself, reports about a post, or corrections to a staff answer can go through the contact page. Posting guidelines explain what belongs here and how moderation works.
Who answers here
Maya Lindqvist
Edits the AI text detection and academic integrity sections
Maya runs the text detection desk. She reads the detector studies as they come out, keeps the accuracy and bias threads current, and writes the answers on how to read a score without over-reading it.
Tomas Reyes
Edits the deepfakes and provenance sections
Tomas covers synthetic media and provenance. He follows the C2PA specifications, watermarking research and the perception studies on synthetic faces, and writes the forum's answers on what people can and cannot verify.