Skip to main content
European Commission logo print header

Anticipatory Learning for Reliable Phishing Prevention

Description du projet

Towards a global dependability and security framework
Keeping a step ahead of the phishers

New software developed by European researchers could go a long way towards preventing criminals from mounting e-mail phishing scams.

Phishing has been one of the major online criminal growth industries of recent years, with billions of scam e-mails sent out to ordinary people around the world.

Often purporting to come from banks and other financial institutions, the e-mails request confirmation of personal account details. Equipped with all of the information they need to get into accounts, the criminals can then strip hard earned savings from them.

Set up in 2006, the AntiPhish project groups together five high-profile partners, each with their own substantial in-house research operations.

The partners include computer software developer Symantec, multinational internet service provider Tiscali, and telecommunications equipment manufacturer Nortel. Fraunhofer-Gesellschaft, the largest organisation for applied research in Europe along with top Belgian university K.U. Leuven are also partners in the effort.

Protecting e-mail globally

The project’s researchers aimed to create new anti-phishing filtering technologies that would help protect and secure the global e-mail communications infrastructure.

That infrastructure has to handle billions of messages a day, a large fraction of them spam and a significant proportion of that is phishing spam.

Because of the huge volume of mail to be checked, and the new tricks spammers are continually using to foil filtering, researchers must make a trade-off between manual effort and effective automatic detection when considering how to improve filtering. Instead of detecting phishing email signatures manually the new software employs trained automatic filters.

Testing in real time

The researchers set out initially to create a series of algorithms – computer programs with specific tasks – under laboratory conditions. They then tested software based on these algorithms in simulated real-world conditions via Symantec’s security network.

After further refinements in the lab, the researchers then tested the software with Tiscali’s users under real-world and real-time conditions. During the Tiscali phase of the trials, part of the company’s huge volume of e-mail traffic was filtered online in real time using specially developed adaptable software.

Anticipating new tricks

These filters are able to identify mail that has similar characteristics to previous phishing attacks and then dump any suspicious ones in a spam folder, an improvement on current technology.

In addition, the software is ‘intelligent’ enough to actually be able to anticipate new types of phishing attacks and block them. This is brand new technology. It means that, as phishing threats continue to adapt quickly to counter-measures, the software is able to keep a step ahead of the con artists.

The technology will add to the protection e-mail users will be able to use in countering phishing attacks aimed at separating them from their money.

One of the most harmful forms of email spam is phishing. Criminals are trying to convince unsuspecting online customers to surrender passwords, account numbers, social security numbers or other personal information. To this end they use spoofed messages which are masqueraded as coming from reputable online businesses such as financial institutions. Phishing has increased enormously over the last months and is a serious threat to global security and economy. The main goal of AntiPhish is to develop anticipatory anti-phishing technologies that help to protect and secure the global email communication infrastructure. With this goal AntiPhish targets exactly strategic objective 2.4.3 which aims at forming scientific, technical and industrial excellence ensuring security, privacy, and trust in complex communication networks and information infrastructures.

The scientific focus of the project is on trainable and adaptive filters that are not only able to identify variations of previous phishing messages, but are capable to anticipate new forms of phishing attacks. Such technology does not exist yet, but would greatly improve all existing methods used in spam and phishing filters. AntiPhish aims at developing filters with unprecedented accuracy which are able to block more than 90% of phishing without blocking more than one legitimate email in a million.

Our consortium comprises reputable research institutions, world leading companies in commercial spam filtering technologies, and a major internet service provider. The research will be driven by the hands-on expertise of our industrial participants, who have long experience with spam fighting on a global scale and will provide huge amounts of real world training data. The AntiPhish project not only aims at developing the filter methodology in a test laboratory setting, but has the explicit goal to implement this technology at internet service providers to be used to filter all email traffic online in real time.

Appel à propositions

FP6-2004-IST-4
Voir d’autres projets de cet appel

Coordinateur

FRAUNHOFER GESELLSCHAFT ZUR FORDERUNG DER ANGEWANDTEN FORSCHUNG EV
Contribution de l’UE
€ 614 442,00
Adresse
HANSASTRASSE 27C
80686 Munchen
Allemagne

Voir sur la carte

Région
Bayern Oberbayern München, Kreisfreie Stadt
Type d’activité
Research Organisations
Liens
Coût total
Aucune donnée

Participants (6)