Artificial Agents and the Law: Some Preliminary Considerations
I am grateful to Concurring Opinions for hosting this online symposium on my book A Legal Theory for Autonomous Artificial Agents. There has already been some discussion here; I’m hoping that once the book has been read and its actual arguments engaged with, we can have a more substantive discussion. (I notice that James Grimmelmann and Sonia Katyal have already posted very thoughtful responses; I intend to respond to those in separate posts later.)
Last week, I spoke on the book at Bard College, to a mixed audience of philosophy, computer science, and cognitive science faculty and students. The question-and-answer session was quite lively and our conversations continued over dinner later. Some of the questions that were directed at me are quite familiar to me by now: Why make any change in the legal status of artificial agents? That is, why elevate them from non-entities in the ontology of the law to the status of legal agents, or possibly even beyond? How can an artificial agent, which lacks the supposedly distinctively-human characteristics of <insert consciousness, free-will, rationality, autonomy, subjectivity, phenomenal experience here> ever be considered an “agent” or a “person”? Aren’t you abusing language when you say that a program or a robot can be attributed knowledge? How can those kinds of things ever ”know” anything? Who is doing the knowing?
I’ll be addressing questions like these and others during this online symposium; for the time being, I’d like to make a couple of general remarks.
The modest changes in legal doctrine proposed in our book are largely driven by two considerations.
First, existent legal doctrine, in a couple of domains, most especially contracting, which kicks off our discussion and serves as the foundations for the eventual development of the book’s thesis, is placed under considerable strain by its current treatment of highly sophisticated artificial agents. We could maintain current contracting doctrines as is (i.e., merely tweak them to accommodate artificial agents without changing their status vis-a-vis contracting) but run the risk of imposing implausible readings of contract theories in doing so. This might be seen as a reasonable price to pay so that we can maintain our intuitions about the kinds of beings some of us take artificial agents to be. I’d suggest this kind of retention of intuitions starts to become increasingly untenable when we see the disparateness in the entities that are placed in the same legal category. (Are autonomous robots really just the same as tools like hammers?). Furthermore, as we argue in Chapter 1 and 2, there is a perfectly coherent path we can take to start to consider such artificial agents as legal agents (perhaps without legal personality at first). This strategy is philosophically and legally coherent and the argument is developed in detail in Chapter 1 and 2. The argument in the latter case suggest that they be considered as legal agents for the purpose of contracting; that in the former lays out a prima facieargument for considering them legal agents. Furthermore, in Chapter 2, we say “The most cogent reason for adopting the agency law approach to artificial agents in the context of the contracting problem is to allow the law to distinguish in a principled way between those contracts entered into by an artificial agent that should bind a principal and those that should not.”
Which brings me to my second point. A change in a legal doctrine can bring about better outcomes. One of the crucial arguments in our Chapter 2, (one I really hope readers engage with) is an assessment in the economic dimension of contracting by artificial agents considered as legal agents. I share the skepticism of those in the legal academy that economic analysis of law not drive all doctrinal changes but in this case, I’d suggest the risk allocation does work out better. As we note “Arguably, agency law principles in the context of contracting are economically efficient in the sense of correctly allocating the risk of erroneous agent behavior on the least-cost avoider (Rasmusen 2004, 369). Therefore, the case for the application of agency doctrine to artificial agents in the contractual context is strengthened if we can show similar considerations apply in the case of artificial agents as do in the case of human agents, so that similar rules of apportioning liability between the principal and the third party should also apply.”
And I think we do.
An even stronger argument can be made when it comes to privacy. In Chapter 3, the dismissal of the Google defense (“if humans don’t read your email, your privacy is not violated”) is enabled precisely by treating artificial agents as legal agents. (This follows on the heel of an analysis of knowledge attribution to artificial agents so that they can be considered legal agents for the purpose of knowledge attribution.)
Much more on this in the next few days.