2026 ASEE Annual Conference & Exposition

The purpose of this WIP research paper is to discuss a pattern in computer science publications that we term ``sidepromising," wherein researchers develop harmful technologies while framing their work in a morally positive light via connections to hypothetical future benefits. First, we discuss the widespread trend of research in robotics being justified with search-and-rescue applications, the ``search" portion of which which may function the same for search-and-surveil or search-and-destroy tasks. Despite a morally positive framing, ostensible robot rescuers often have more obvious applications and funding relationships to systems of oppression like the military industrial complex. Second, we discuss machine learning research, which is widely justified with claims about superhuman abilities to solve our most pressing problems (e.g., the ``AI will cure cancer and solve politics” arguments). We argue that it is disingenuous to focus only on (often unrealized) benefits to motivate research that contributes to killing and otherwise harming people, accelerating climate change, and empowering fascist repression. We then present early procedural, technological, and social strategies to mitigate sidepromising in academic publications.