Experimental usage of AI brain-computer interfaces: computerized errors, side-effects, and alteration of personality
Stevens, I and Gilbert, F, Experimental usage of AI brain-computer interfaces: computerized errors, side-effects, and alteration of personality, Ethics of Medical Innovation, Experimentation, and Enhancement in Military and Humanitarian Contexts, Springer International Publishing, D Messelken and D Winkler (ed), Cham, Switzerland, pp. 195-209. ISBN 9783030363192 (2020) [Research Book Chapter]
The US Defense Advanced Research Projects Agency (DARPA) is currently funding experimental trials testing in human novel medical brain implants operated by Artificial Intelligence (AI). The purpose of this chapter is to explore some ethical issues related to the experimental use of these invasive AI-controlled brain devices, in particular, deleterious phenomenological effects these devices may have on a patientís personality and/or sense of self (i.e. patients suffering from postoperative self-estrangement despite symptom reductions). The evolution of these devices from open-looped stimulation to closed-loop personalized AI-controlled stimulation raises many safety concerns that may exacerbate these ethical issues. This new AI-controlled approach is unlike previous open-loop methods (i.e. traditional deep brain stimulation); the AI-tailored made frequency stimulation schedule depends on the computational measurement of patientsí brain states which fluctuates from patient to patient. Hence no universal safety standard and the potential for computational error resulting in plausible deleterious effects on a patientís personality. The aim of this chapter is to explore how closed-loop stimulation undermines safety standards and results in skewed risk assessments for complex phenomenon such as a patientís personality, but as well autonomy.
Research Book Chapter
artificial intelligence, autonomy, brain-computer interfaces, experimental trial, implant, military medical obligation