Liability for autonomous decoders
The problem of allocating legal responsibility when a semi-autonomous BCI produces an action or output the user did not consciously intend. If a decoder misclassifies an intention and moves a prosthetic or wheelchair, or a closed-loop stimulator triggers a behavioral change, who is liable — the manufacturer, the clinician, the user, or no one?
Scholars call the risk a 'responsibility gap': autonomy and continual adaptation can break the chain that links a human decision to an outcome. Existing tools include product-liability law — the EU's revised Product Liability Directive of 2024 explicitly covers software and AI — and medical-malpractice doctrine, but questions of user agency, foreseeability, and the device's own learning strain them.
The proposed EU AI Liability Directive was withdrawn in early 2025, leaving general product-liability and negligence law to carry the load. A distinctive neurotech twist is agency itself: because the device operates on the very signals of intention, it can be genuinely unclear whether an action was the user's or the decoder's.