Artificial intelligence (AI) systems connected to sensor-laden devicesare becoming pervasive, which has notable implications for a range of AIrisks, including to privacy, the environment, autonomy and more. Thereis therefore a growing need for increased accountability around theresponsible development and deployment of these technologies. Herewe highlight the dimensions of risk associated with AI systems that arisefrom the material affordances of sensors and their underlying calculativemodels. We propose a sensor-sensitive framework for diagnosing theserisks, complementing existing approaches such as the US National Institute of Standards and Technology AI Risk Management Framework and the European Union AI Act, and discuss its implementation. We concludeby advocating for increased attention to the materiality of algorithmicsystems, and of on-device AI sensors in particular, and highlight the needfor development of a sensor design paradigm that empowers users andcommunities and leads to a future of increased fairness, accountability andtransparency.
This paper introduces a standard datasheet template for ML sensors and discusses its essential components inluding: the system's hardware, ML model and dataset attributes, end-to-end performance metrics, and environmental impact. We provide an example datasheet for our own ML sensor and discuss each section in detail. We highlight how these datasheets can facilitate better understanding and utilization of sensor data in ML applications, and we provide objective measures upon which system performance can be evaluated and compared.
In this viewpoint we propose the ML sensor: a logical framework for developing ML-enabled embedded systems which empowers end users through its privacy-by-design approach. By limiting the data interface, the ML sensor paradigm helps ensure that no user information can be extracted beyond the scope of the sensor’s functionality. Our proposed definition is as follows: An ML sensor is a self-contained, embedded system that utilizes machine learning to process sensor data on-device – logically decoupling data computation from the main application processor and limiting the data access of the wider system to high-level ML model outputs.