How AI helps check it
Messages
We taught the computer using over 1,100 real messages received by people in Malaysia, both scams and ordinary ones. It learned how scammers write: rushing you, promising prizes, asking for account numbers.
Links
We look at how the address is written, for example a bank's name placed where it does not belong. We never open the link.
Reported links
We also ask two security services, Google Safe Browsing and VirusTotal, whether the link has been reported before.
Official addresses
Real bank and government addresses are checked against official lists. This part is not AI. It is a direct check.
How we keep you safe
- We never open the link you paste.
- We remove phone numbers, IC numbers, emails and account numbers before saving anything.
- Every report is read by a person first. Nobody can teach ScamSense that a scam is safe just by reporting it many times.
What ScamSense cannot do
- It can be wrong. Sometimes it is extra careful, and even a real website can get a warning.
- It never sees the page itself, only the address.