CatchFish

AI image generators have telltale signs. Look at ears — often asymmetric or malformed. Hair is impossibly perfect. Jewelry and glasses have loops that don't close. Backgrounds look dream-like with objects that don't make sense.

Eyes look "glassy" with identical catchlights in both eyes. Real photos rarely look this symmetric. Check if the photo exists elsewhere using Google Images, TinEye, or Yandex reverse image search.

Free tools: hivemoderation.com and aiornot.com can scan suspicious images.

Watch for these scripted phrases — scammers reuse them constantly:

If you already clicked something and it's too late for "don't click" advice, here's what to do next, in order.

1. Don't enter anything: If a page opened asking for a password, card number, or personal info, close it without typing anything in.

2. If you entered a password: Change it immediately, from a different device if possible, and change it anywhere else you reused that same password.

3. If you entered financial info: Call your bank or card issuer's fraud line right away — the window to reverse a wire transfer or freeze a gift card is measured in hours, not days.

4. If your Social Security number was involved: Contact Equifax, Experian, and TransUnion to place a fraud alert or credit freeze.

5. Report it: File at ReportFraud.ftc.gov and IC3.gov (note: IC3's form requires a desktop or laptop browser, not mobile). Also forward the email to reportphishing@apwg.org.

Save their photo, then go to images.google.com → click the camera icon → upload the photo. If the same face appears under different names or on stock photo sites, it's likely stolen.

Also try tineye.com and yandex.com/images — Yandex is especially powerful for finding Instagram models whose photos are commonly stolen.

Scammers frequently steal photos from military personnel, nurses, and fitness influencers who have large public photo libraries.

If someone's using your photo without permission, a DMCA takedown notice is one common way to get it removed — but CatchFish doesn't file these for you or give legal advice. Here's what to expect if you look into it yourself.

What it is: A formal notice you (as the copyright owner) send to a website or its host, asking them to remove content that infringes your copyright. Most platforms have a dedicated process for this — look for "Report Copyright Infringement" or a DMCA contact in their footer or help center.

What it generally requires: Identifying the specific content and where it appears, a statement that you're the rights holder (or authorized to act for them), your contact information, and a good-faith statement that the use isn't authorized. Requirements vary slightly by platform.

What a lawyer typically helps with: Confirming you actually hold a valid claim, drafting the notice correctly, handling it if the other side disputes it ("counter-notice"), and advising on next steps if the platform doesn't act.

When this stops being DIY: If the situation involves harassment, threats, blackmail, or identity theft, that's no longer just a copyright issue — contact a lawyer or local law enforcement directly rather than relying on a takedown notice alone.

Ask hyper-local questions: "What's your favorite restaurant near [their stated neighborhood]?" A real person answers specifically. Scammers deflect or give vague answers.

Reference something specific from earlier in your conversation. Bots and script-readers often miss this context.

Insist on a live video call. Make them hold up a specific number of fingers or wave. Scammers always have an excuse — bad connection, broken camera, traveling, shy.

Paper Check doesn't run a fingerprint scan or compare against a database of known AI writing. It's Claude reading the paper and making a judgment call, the same way an attentive instructor might notice something feels off — based on patterns, not proof.

What it actually looks for: generic phrasing that could apply to any topic, a lack of specific supporting detail, repetitive sentence structure, and whether the writing lines up with the stated textbook and course level.

What this means in practice: no AI-detection tool, including this one, can tell you with certainty whether a paper was AI-written. Every tool in this category has real false-positive rates — and multiple studies have found they disproportionately flag writing by non-native English speakers, even when it's entirely their own work.

Why we still built it: a pattern worth a follow-up conversation with a student is still useful information, even when it isn't proof. Paper Check is designed to point you toward a conversation, never to replace one.

"Pig butchering" is a long-con romance scam where the scammer builds real emotional trust over weeks before introducing a crypto "investment." They "fatten" you with affection before "slaughtering" your savings.

Red flags: They casually mention crypto profits. They offer to help you invest "risk-free." They show you a trading platform you've never heard of — these apps are fake, designed to show phony gains until you try to withdraw.

Rule: if anyone you've never met in person mentions cryptocurrency or investing — end contact immediately and report to reportfraud.ftc.gov.