The Developer Factor in Software Privacy

Abstract

Computer programming operates and controls our personal devices, cars, and infrastructures. These programs are written by software developers who use tools, software development platforms, and online resources to build systems used by billions of people. As we move towards societies that rely on computer programs, the need for private and secure systems increases. Developers, the workforce behind the data economy, impact these systems' privacy, and consequently, the users and society. Therefore, understanding the developer factor in software privacy provides invaluable inputs to software companies, regulators, and tool builders.
This thesis includes six research papers that look at the developer factor in software privacy. We find that developers impact software privacy and are also influenced by external entities such as tools, platforms, academia, and regulators. For example, changes in regulations create challenges and hurdles for developers, such as creating privacy policies, managing permissions, and keeping user data private and secure. Developers interactions with tools and software development platforms, shape their understanding of what privacy means, such as consent and access control. Presentation of privacy information and options on platforms also heavily impact developers' decisions for their users' privacy, and platforms may sometimes nudge developers into sharing more of their users' data by using design (dark) patterns.
Other places developers learn about privacy include universities, though they may not learn how to include privacy in software. Some organisations are making efforts to champion privacy as a concept inside development teams, and we find that this direction shows promise as it gives developers direct access to a champion who cares about privacy. However, we also find that their organisation or the wider community may not always support these privacy champions. Privacy champions face an uphill battle to counter many of the same privacy misconceptions seen in the general population, such as the `I’ve got nothing to hide' attitude.
Overall, I find that research in developer-centred privacy is improving and that many of the approaches tried show promise. However, future work is still needed to understand how to best present privacy concepts to developers in ways that support their existing workflows.

Publication
The Universtiy of Edinburgh