Privacy: A Surmountable Challenge for Computer Vision

Ethical concerns about privacy come with the territory of computer vision—after all, it’s hard to deny the privacy implications when considering the digital images or videos that are the bread and butter of these applications. But privacy concerns need not be a roadblock that stands in the way of technological innovation. While the idea of “informed consent” has been a popular method of preserving privacy in some domains, it may not always be either possible or desirable in the context of emerging technologies like computer vision. In this talk, we will consider a more flexible approach to preserving privacy that pays special attention to the context that a device is being designed for. In addition, we will review several examples that demonstrate how small changes to the design of technology can make a big difference when it comes to preserving privacy.