ScamDoc: now available as a browser extension!

Last year, I worked with Anthony Legros and Jean-Baptiste Boisseau on an AI to detect fraudulent emails or websites. This 2-month internship was a great opportunity to discover the world of online scams and to apply my skills in machine learning to a concrete problem: how can we use AI tools to identify email or website addresses that are scams?


Anthony Legros and Jean-Baptiste Boisseau have been fighting online scams for years in particular through their participative website www.signal-arnaques.com. They’ve now added some machine learning to their toolbox to automatically evaluate “digital identities” trust. A year ago, we collaborated on the design and the implementation of their free tool to check for scams online: ScamDoc.

A quick peek at ScamDoc

In a nutshell, ScamDoc is a free online service that allows you to easily enter an email or a website address and automatically get a score for it that predicts how “dangerous” it is: the closer to zero, the better! This website relies on ScamPredictor, an ensemble of AIs that I was tasked with conceiving and coding up during my internship with the help of Jean-Baptiste Boisseau.

ScamDoc uses AI to predict a confidence score for email or website addresses and helps you to know whether you should trust it or not!

The service is absolutely free of charge and can be used as many times as you wish, however there is also a pro (non-free) API for companies to access ScamPredictor directly from anywhere.

The tool can be applied both to emails or websites – but it doesn’t read the content, it just analyzes the address itself and applies various methods to predict a score (neural networks, Bayesian filters…).

If you want to learn more about how ScamDoc works, don’t hesitate to check out the article I wrote about the project several months ago.

In only a year, ScamDoc has already had lots of really positive reviews and it has been upvoted by the community, for example by the youtuber Sandoz (video in French).

What’s new?

Jean-Baptiste just released a new very neat tool to make it even easier to benefit from ScamDoc’s advice: a browser extension available for Chrome (and Brave, since it’s the same engine under the hood!) and Firefox users.

Once you’ve installed the add-on, it will automatically check the websites that you browse and, if ScamDoc predicts a really “scamous” score, it will warn you with a little alert panel. You can still go to the site of course but you’ll be aware it looks dangerous.

If you go to a website that ScamDoc assesses as dangerous, it will warn you with an alert panel like this one.

You’ll even have a direct link to a page that shows how ScamDoc assessed the harmfulness of the address; there are various criteria such as: how old the WHOIS of the website is, how long it will last, in which country its registrar is…

You can see the detailed score of the website and why it has this score by going to scamdoc.com.

All in all, I’ve found it to be really useful. It’s super easy to install and use. The only small issue I’ve had is that most recently created sites trigger the plugin even though they may actually be trustworthy. But, once again, the extension doesn’t impose anything, so you’re free to just disregard the alert. And in some cases, it’s quite interesting to have it pop up and suddenly get way more cautious.

We sometimes forget to stay alert and the ScamDoc extension is a good reminder that there are fraudulent well-designed websites, too.

What else?

The service is continuously being improved by the team and should soon integrate IP adresses and phone numbers as well. Keep in touch for more info on the new features!

Also, if you want to stay up to-date with the world of scams, I recommend you follow Signal Arnaques/ScamDoc either on their blog (in French) or their Twitter: they regularly publish lots of articles on various scam-related topics (how to use Google Alerts to check your e-reputation, how to identify and avoid common scammer tricks…).

References
  1. Signal Arnaques’ blog: https://info.signal-arnaques.com/
  2. ScamDoc’s website: https://www.scamdoc.com/
  3. ScamDoc’s Twitter account: https://twitter.com/scamdoc
  4. ScamDoc browser extension on Chrome/Brave: https://chrome.google.com/webstore/detail/scamdoc/fojbeoliffkbkfgaehkoeeihphibcjdk?hl=fr
  5. ScamDoc browser extension on Firefox: https://addons.mozilla.org/fr/firefox/addon/scamdoc/

Leave a Reply

Your email address will not be published. Required fields are marked *