How IGNIS by Saphira Uses AI
Last updated: [DATE]
Product site: https://askignis.com/ai-disclosure
The short version
- Your scam verdicts are made by computer AI. [Most/All] verdicts are not checked by a person.
- AI can be wrong. It can miss a scam or wrongly flag a real message.
- Always double-check before sending money or personal information. Call the company using a number you already trust.
How it works
- You send us a text, email, screenshot, or link.
- Our system and an AI model from [AI VENDOR(S)] read it and look for warning signs, like urgent threats, odd links, or requests for gift cards.
- We send back a verdict, such as "Likely Scam," "Be Careful," or "Looks OK," with a short reason.
- The family Account Holder may get an alert.
Do people review it?
[BUSINESS DECISION — choose one:] - Option A: "No person reviews most verdicts. Staff may look at a message only if you ask for help, to fix a problem, or to keep the service safe." - Option B: "A trained team member reviews messages marked uncertain or high-risk (e.g., large money requests)."
Is my data used to train AI?
- Our AI vendors do not use your messages to train their models, under our contracts and settings.
- IGNIS by Saphira [does not / with your opt-in only] use your message content to improve our own system. We may use de-identified scam patterns.
What we never do with AI
- Make decisions about your credit, insurance, jobs, or housing.
- Contact anyone or move money for you.
- Pretend to be a human. If you ask, we'll tell you it's AI.
Getting help
Not sure about a verdict? Reply "HUMAN" [if offered] or email [PLACEHOLDER: support email]. In an emergency, call 911. If money was sent, call your bank right away and report at ReportFraud.ftc.gov.
[LAWYER: check state AI disclosure laws (e.g., Utah Artificial Intelligence Policy Act; California bot disclosure law, Bus. & Prof. Code §17940 et seq.; Colorado AI Act) and FTC guidance on AI claims — avoid overstating accuracy in marketing (FTC Act §5).]