Title: Ending EULA Roulette
Speaker: Ms.Rawan Alabdulrahman, Software Engineering Department.
End-user license agreements (EULAs) are problematic. Users tend to not read them; they assume EULAs are unenforceable, but actually they are spinning a roulette wheel. They could trap themselves into legal complications because courts are perfectly capable of enforcing them. The EULA system, we argue, violates the informed consent contracting principle, and does not follow the equality of bargaining power principle. These two problems are not typical software engineering ones, but the dilemma discussed here can only be seen and tackled by software engineers. The research’s ethos and aims are to follow the University College London (UCL) Engineering ambition to “Change the World”. We do that by trying to break the vicious circle that forces users to consent to risky EULAs through a simple button-click and that keeps individual users at the wrong end of the bargaining process with large corporations. To this end, this work will introduce the “Yes-Man” tool that tackles EULA system by clicking the EULA-associated “Accept” button automatically. It does this by operating as a virus. As such, the Yes-Man tool is not trying to act maliciously, it is to show that the EULA contract does not offer a solid informed consent process. Our work will also create a new legal service “EqualisingEULA” that provides a EULA related information to users. The purpose is to empower the end-user party towards restoring the balance in the EULA system. Both contributions are based on the “ISeeEULA” tool. It can automatically identify windows-based EULA via two mechanisms: EULA’s GUI features recognition and EULA’s agreement ML classification. The ML classification utilises the TF-IDF method to drive features and a voting system classifier to classify an input EULA. Since both Yes-Man and EqualisingEULA rely on ISeeEULA, we empirically validated and evaluated our ML approach on a corpus of 116 EULAs and 113 non-EULAs (i.e. other legal contracts). The results showed high accuracy, good precision, and recall, thus confirming the feasibility of our approach.
Date/Time: Tuesday, Jan 17, 2017 at 12-1pm
Location: Khadija Auditorium, 6G41/6F49 in Building 6 (Broadcast to Room 2084 in CCIS male side)
[Back to the Events page]