Brain Computer Interfaces: Privacy, Ethics, and Policy

Thesis (Ph.D.)--University of Washington, 2019The ability to record and access neural signals is expanding, primarily driven by commercial entities developing non-invasive headsets that can control digital devices using a person’s neural electrical signals. This is an excitingly novel modality for interacting with technology that will allow companies to profit from the sale of these devices. But hardware isn’t the only thing of value in this exchange: companies also get access to the neural information of users. While some may choose to only use these signals for control purposes, this represents a new way that information can be elicited from someone, knowingly or unknowingly. Some may think nothing of revealing their preferred cereal or the kind of car they drive, because data brokers already collect hundreds of data points about us and interpolate many more. But what about the things that you wouldn’t want to share, like a credit card number, whether you committed a crime, or perhaps even your sexual orientation? This is just one example of the many questions about privacy and perceptions of neurally-derived information that are becoming apparently for this emerging technology. These recordings of the electrical activity in the brain fall into an unregulated area of law: while they are medically-based, they are obtained for non-medical purposes and thus do not fall under the jurisdiction of existing legislation to protect health information (HIPAA, Health Insurance Portability and Accountability Act). A general absence of federal data privacy legislation and overall limited enforcement mechanisms on behalf of the Federal Trade Commission (FTC) means that new sources of consumer information are at risk of being exploited like other kinds of data are currently (e.g. from social media). Companies may claim that they are not selling the data they have on you, or that their practices are clearly outlined in the terms of service you accepted, but this only belies how opaque this particularly industry has become; consumers cannot say what happens to their information, but they inherently believe that it is only being used for the things they think it is (not what is actually happening to it). While lawmakers act reactively to address these kinds of known issues, it is possible to be proactive in the development and regulation of nascent technologies with commercial potential like neurally-controlled devices. This dissertation describes a multi-disciplinary research approach to addressing the issues that arise from emerging technologies that rely on access to neural signals. It begins with a literature review of the techniques used to elicit information, as well as prior work. Then, experimental results will be presented that demonstrate the kinds of information that can come from real-time analysis that does not rely on prior classifier data. The following chapter discusses the philosophical foundation of neural privacy and presents the results of a neuroethics survey that demonstrates how individuals perceive the privacy of their neural information. Finally, the existing regulatory landscape will be discussed with respect to how neural information can be better regulated; this is done based on responses to policy-based questions from the neuroethics survey