Inspiration

Scammers send hundreds of messages today, typically offering something with great value such as free technology under the disguise of obtaining personal information from unsuspecting victims to take advantage of.

Normally the best defense against these scammers is awareness and choosing not to engage with any potential scam messages they put out. While this is plausible and protects individuals from scams, the scammer doesn't stop here and can move on to the next person to target with the same message. Often times, scammers get away with their actions and nothing they reveal about themselves (such as payment information and emails) are ever recorded for public safety.

We wanted to make a change against this normal of scammers getting away with absolutely no consequence to their action. What if instead of ignoring them, someone talked back? An AI chatbot can keep the scammer engaged and waste their time along with fishing for their personal payment and email information to build a report that can be used to be sent to fraud-report services and platform moderators for further action.

We were inspired to be part of the scambaiting community and automate what they do to address the hundreds of scam messages that are often overlooked and tedious to address, so anyone can fight back against scams.

What it does

Our project is a Discord bot that talks to scammers so other individuals don't have to. When a suspected scammer sends a DM, the bot engages them by starting conversation and fishing for details about the scam that often leads to payments, emails, crypto, and more.

After evidence is collected, the bot makes a report with the included transcript and the gathered information about the scammer when the scammer stops responding for a set amount of time. The bot then emails this report after.

How we built it

We utilized Gemini for vibe coding and also hand coded most of the features.

Challenges we ran into

Challenges we ran into include tweaking the model to sound more human and getting it to send post-interaction reports via email.

Accomplishments that we're proud of

We made the bot sound sufficiently human and built so that it doesn't really give up on fishing for information, keeping the scammer around even when they get frusturated or suspicious. This is a skill that is extremely hard to replicate, and one that we managed to implemented through patterns fed into the bot.

What we learned

We learned many things about how scammers function and what sort of language keeps them around. Working on Gemini showed us how a product can be structured to incorporate AI efficiently to prioritize feature planning and scope of the project.

What's next for Scammer Whammer

In the future, we want to implement voice chat and/or phone calls. This would be extremely hard to do, especially because AI is currently not advanced enough to completely replicate how a normal person would talk and could easily be recognized by those familiar with AI. We also would want to expand this project to have more human-like features like perhaps sending pictures for further successful scam-baiting.