Legal and Social Implications of Predictive Brain Machine Interfaces: Duty of Care, Negligence, and Criminal Responsibility

Neuroscience has taken us even beyond the ability to predict the onset of a seizure. Recent research has shown that brain–machine interfaces (BMIs), particularly those associated with neuroprosthetics, and especially those situated in the brain’s posterior parietal cortex (PPC), can tap into subconscious thought (Thomson 2015). This technology has the ability to even predict and trigger user actions— ostensibly to prevent lurching, jerky, and/or delayed prosthesis motions—even before that user consciously knows that he or she is going to make them (Aflalo et al. 2015). These and other BMI systems raise legal and ethical concerns perhaps more disquieting than patient autonomy and privacy: in particular, the very practical, and likely soon-to-be-pressing, issues relating to the ethical and legal duties associated with tort and criminal law. For example, legal and regulatory systems will need to deal with BMIassociated human enhancement technologies such as exoskeletons (AAS EuakALERT 2015) or prosthetics (Jacob 2015) that, as a result of unanticipated subconscious thought, cause harm to their users, others, or property. Even more nefarious, the BMI could be hacked such that a set of actions is maliciously forced on the user. This isn’t just a far dystopian future: Although BMIs still suffer from high levels of false positives and false negatives, the microchips in the aforementioned PPC experiments are now commercially available and they are even approved for human use (Pruszynski and Diedrichsen 2015). And, Panasonic, for example, promises relatively affordable consumer exoskeletons by the end of 2015 (Russon 2015). It’s not just the use of PPC BMI implants that are potentially problematic. With many prosthetics incorporating artificial intelligence technologies (Rupp et al. 2014), we will likely see a further muddying of the waters in deciphering who or what actually caused the resulting offensive actions, an important legal determination. These issues are likely not going to be limited to just the disabled community: Otherwise abled individuals could use this technology in the military, heavy industry, and other areas. BMIs could fall within the nascent field of robot law— like robots, they obtain information in the form of electrochemical neuronal signals, they interpret these signals through complex algorithms, and then they output a signal to an effector that may result in a perceived action. However, in contrast to other areas of robotics law, PPC BMIs raise additional and unique concerns: In addition to all the standard relevant stakeholders including the programmers and manufacturers, the law must account for the confounding additional element of the potentially subconscious integrated user who introduces substantial uncertainties typically associated with all biological systems, uncertainties that are further confounded by potentially unpredictable artificial intelligence. U.S. tort and criminal law, although at one point doctrinally aligned (Seipp 1996), differ significantly in a