Do Robots need their own Regulatory body?

Dr Aaron Mannes’ paper ‘Institutional Options for Robot Governance’, examines the possible options for how to regulate robotics at a federal level. One option would be to set up a completely new agency or bureaucracy that deals exclusively with robotics. It would be a centre of robotic expertise, incorporating a high degree of legal and political expertise.

Is this necessarily the best approach though? Would it be better to consider altering an existing agency in some way instead?

The major appeal of forming a new agency is that it would be able to bring in new ideas that are relatively unconstrained by past ideas and perspectives; it could start with a clean slate. Making these decisions now, at this early stage, is important to ensure that the best possible future for robotics. A comparison is made to the internet where early decisions, such as granting first amendment rights across the platform, helped ensure the new technology would prosper. Robotics has a direct, physical presence in the world as opposed to being used to relay information. This allowed the development of the internet to be dealt with in a slightly more hands-off manner, something which could not be replicated for robotics.

Establishing a new agency presents a series of difficulties. Firstly, on whose authority would it be established? A president could legally set up a new robotics agency, but this route would allow Congress the opportunity to defund it, and it runs the risk of being dis-established due to the slow nature of an agency developing its specific programs. Then there are practical, logistical concerns. A new agency will need office space, furniture, specialist equipment and facilities as well as the right staff.

The potential for turf wars with other agencies is also a concern. If it is not given suitable regulation and a formalised statutory role, other agencies can choose to ignore it; and if it is given a formal role, other agencies might push back at it, seeing it as a threat to their own autonomy. Cooperation with other agencies is of paramount importance, if the new body is to succeed.

Appointing a current agency to act as a lead agency for robotics is one such method of operating within a pre-existing framework, developing and expanding its relevant capabilities. The principal advantage of this route is that the infrastructure of the agency will already be in place, cutting down the development time. Ease of setting up is an attractive feature but it does come with its own issues. Dilution of the existing agency’s primary function would be a concern and could lead to a lack of enthusiasm for its new department.

The issue of which agency should take the lead is also not clear-cut. NASA, given that it already creates robots, is considered to be a frontrunner (perhaps becoming NASRA). But the lack of expertise with the legal and policy aspects involved with robotics means that they might be less interested in pursuing this new direction.

A series of other, smaller options that act as half-measures include setting up working groups inside agencies, Congressional committees, appointing a ‘Czar’ as a central public figure to tackle the issue head-on, or ‘blue ribbon’ committees that are intended to be temporary strategic fixes. While these options lack the authority of setting up an agency from scratch or appointing a lead agency, they have the benefit of immediacy, requiring considerably less effort in the initial stages.

Another option that is worth discussing is simply to do nothing. Allow the existing agencies the opportunity to adapt and to learn, they all have to go through inevitable periods of change and innovation.

Building and modifying bureaucracies is like constructing robots. There will be trade-offs in the design, and a certain amount of unpredictable, emergent behaviour is to be expected.