This is an idea to fight scamming that is becoming more frequent in Jakarta using face identifying to easily know if the suspect is scamming or actually in need.

Disclaimer

No AI is used on this article nor planned to be used for the app that will be implemented, vibe-coding included.

Background

Scamming become more and more frequent in Jakarta, Indonesia. The modus is that the suspect acts as if he/she need some fund, relatively small to a passerby, to be able to go home.

The common mental respond, including me, is “if he/she is really need help, that’s good, if not, I’m fine either way, it’s just IDR x”. But doing so will encourage more scammer to do that, since the victim is okay with giving money.

But I don’t want it, let’s make that person suffer.

Proposal: GetContact but for peoples’ face

Get contact is an app that stores phone numbers and label them if they’re malicious, or clean. They also put the tags related to the phone number. These tags are what the GetContact users’ label/name the phone number in their phone.

How about we use that but for people’s face. People could just scan the potential scammer’s face, and see if it’s a scam, or they’re clear.

We will need to be able to ID each person’s face, store it in the database. If it’s reported that it’s a scam, the next person could see that the person is a scammer, and will not fall into the trap.

Technologies Involved

1. Face ID technology To easily put an ID to each person’s face. It’s an existing technology, we could just utilize it.

2. Web app Users could easily access the app via web, and no need to download a one-time only app.

Initial Data

To reduce the first users to have to be scammed. We could use existing social media data. I see in Threads or Instagram that someone reported that this person (with a selfie) is suspected to be a scammer with this and that chronology. The data is there, and we could use it, we just need to know how to harvest and utilize them.

FAQs

1. What about privacy? If the suspect is in need of help, well, he/she must comply to our request to take a picture of their face. Legally, I don’t know.

2. What if it’s abused/misused? We will use layers of verifications and limits to drop and reduce potential malicious actors:

   1. We require the user to logged in into an email or phone verified account.
   2. We should limit each user to be able to scan or report to a limit per day. Who got scammed more than once a day? A very unlucky person I indeed.
   3. Data-science backed human in the loop. Not sure how, but maybe we could have admins that learns the data and come with an idea of which users and/or pattern are doing malicious activities, such as multiple account being created in a relatively close time and reporting to the limit of each account.
   4. User verification and scoring algorithm. Each user could be more and more trusted and some could be suspected of abuse. How? Not sure yet at the moment.