Current discussion of computational law applications focuses on a human-to-application relationship; i.e., how humans use these applications for various ends. But this is just one configuration. These applications could also play an important and useful role in an application-to-machine (A2M) configuration.

Moving this idea to the context of connected and autonomous vehicles (expanding the term to drones, cars, ships, etc.), or “CAVs,” the computational law application can be seen taking on the role of providing legal advice to the CAV, not the human operator. Why does the CAV need legal advice? Pretty much for the same reason a human operator does. To comply with the law. If it is compliant, it can be deemed “trustworthy.”

The A2M setting shifts the focus (though not necessarily the burden and liability) of legal compliance from a human operator to the CAV. The A2M computational law application consistently analyzes input from various sensors in the CAV (GPS, speedometer, etc.) and outputs relevant operational instructions to it. As the CAV travels through multiple jurisdictions, for example, its operation can be dynamically modified to ensure it consistently remains legally compliant. At the most rudimentary level, this output could manifest in restricting the CAV’s speed to comply with the local limit during rain or directing it move to another lane because an emergency vehicle was detected. More complex scenarios may require instructing the CAV to shift, if only temporarily, from a manual or semi-autonomous mode to fully manual, or fully autonomous.

Computational law A2M applications are conducive to building a compliance-by-design CAV. It coalesces with other “by-design” development principles, such as security and privacy by design; it promotes the existence of trustworthy CAVs. A2M is also a machine learning design paradigm. A2M-enabled CAVs can learn and adopt behavior that (also) syncs with what constitutes as “optimal” in a given jurisdiction. All of this learning can be shared with other CAVs via relevant ontologies, such as the transportation-centric ontology I have written about before here. This is valuable: the more the A2M-enabled CAVs learn, the more trustworthy the entire CAV ecosystem can become.

***Postscript***

June 2, 2021: Qualcomm and the city of Peachtree Corners in Georgia are partnering in a Cellular to Vehicle to Everything (C-V2X) rollout. The effort is designed to promote smart city planning and will include, among other technologies, Vehicle to Infrastructure (V2I) testing. The C-V2X should be viewed as an umbrella framework under which V2I, A2M (discussed above) and other sub-frameworks operate. Each of these sub-frameworks plays an important role in powering, not only the smart city, but also any environment in which CAVs operate.

June 27, 2019: The WSJ reports on HP’s patent applications that describe methods for making IoT smarter by combining blockchain and AI. HP’s “swarm” approach can also be useful for building the transportation-centric ontology. This approach is also reminiscent of the neural network “capsule network” explained in Sabour, Frosst and Hinton’s “Dynamic Routing between Capsules” article, discussed in my Fractal Disambiguation for AI post.

May 24, 2019: Filings of computer vision patent applications have moved into a steep upward trajectory since around 2012, far outpacing any other for AI related applications. This data indicates (among other things) that the industry focus is on cognitive AI applications (e.g., neural networks), a trend which can also promote the growth of computational law A2M applications, and not just in CAV settings.