AI Catfish Dating Scam Defense Lab | Forensic Scanner & Safety Tool

EXIF: STRIPPED ⚠ Pupil Asymmetry & Earring Drift

Sophia Vance

Crypto Lifestyle & Fine Art • Milan / NYC

CRITICAL RISK

ID: verge_ai_scam_01

"Curating decentralized web3 galleries & high-frequency algorithmic aesthetics. Looking for a genuine connection with someone driven and spiritually aligned."

Dating Match: 12 mins ago

Account Age: 2 days (Fresh Sybil)

Triage Chat Feed (App Internal Log)

Heuristic Engine Active

Hey! So glad we matched. Your bio caught my eye. Are you still using WhatsApp? It's easier to share my gallery links there ✨

10:42 AM Flag: App Disintermediation Trap

Execute Defensive Counter-Challenge:

Synthetic Forensic Checklist

Identify all 4 critical automated markers

0 / 4 Detected

AI-Generated Synthetic Iris Inconsistency +25 Pts

Polygon pupil morphology, mismatched reflections, and earring drift on profile render.

Accelerated Romantic Attachment +25 Pts

Unnatural emotional cadence, excessive flattery within 4 initial exchanges (Love-Bombing Bot).

Immediate Off-Platform Migration Pivot +25 Pts

Immediate steering toward encrypted WhatsApp or Telegram to bypass dating app safety bans.

Subtle Crypto / Wealth Signaling +25 Pts

Background high-roller imagery and crypto gallery framing preparing pig-butchering pivot.

Defensive Threat Assessment

Scam Confidence Index: 0% (Inconclusive)

Playbook Match: Pending Triage

Authentication Check: Unchallenged

Action Recommendation: Gather Forensic Evidence

A local checklist lab assigns scores and scripted responses

Your browser does not support video.

Read the explanation

The saved lab has four boolean flags: iris, attachment, migration and crypto. Each checked flag adds twenty five points, so one gives twenty five, two fifty, three seventy five and four one hundred. At three pixels per point each contribution spans seventy five pixels and all four adjacent segments span three hundred. The scan button simply sets every flag true; it does not inspect a photograph, read a real conversation, run a model or discover a scam. The four flags are manual or assigned local state. The highest label says confirmed AI catfish, but that wording is not supported by independent identity or fraud verification. This explanation preserves software behavior without repeating the claim as a real classification. The visible score labels zero inconclusive, up to fifty moderate, seventy five high and one hundred critical. The JSON report instead calls three or more flags critical and every lower count elevated. Thus three flags gives a high label on screen but critical in the report; zero is inconclusive on screen but elevated in the report. At three pixels per point the UI critical boundary one hundred measures three hundred, report critical boundary seventy five two hundred twenty five and their difference twenty five seventy five. These branches are separate naming rules, not different scientific confidence levels. Actual exports carry the checked flag descriptions, persona fixture and challenge-status strings. A checkbox or report string is not independent forensic evidence. Each of three challenge buttons sets its executed boolean immediately, appends a local user line and schedules one fixed local reply after seven hundred milliseconds. The JSON exporter therefore marks gesture deflected as soon as it is clicked, even before that reply occurs. At point two pixels per millisecond immediate zero has zero width, reply delay seven hundred one hundred forty and their difference one hundred forty. Repeating the same challenge is blocked by its boolean. Reset clears booleans and extra chat but never cancels pending reply timers, so a scheduled reply can arrive after reset. The block button only sets a local flag and appends a system line; no platform ban or report is sent, and that flag is absent from exported report fields. Native controls and JSON can work despite unavailable Tailwind styling. No person, dating service, live camera, safety backend or real message is accessed.