Siri, Who's the Boss? For Whom Do Intelligent Agents Work?

Intelligent agents are here. They exist not only in our smartphones (Apple’s Siri), but increasingly, in our homes (Amazon’s Alexa) and vehicles – consider the workings of the semi-self-driving Tesla. Not too long ago, the view that computers and computer programs were mere tools of the person using them was seen as incontestable. But increasingly, policymakers are confronting the possibility that computer programs may develop the capacity to act as agents in the legal sense. This Article confronts three different established approaches by which to answer a potentially very important legal question: For whom do intelligent agents work? The answer to this question may affect liability under a variety of legal regimes, including contract, tort and competition law, among other areas. First, contract law might answer that intelligent agents’ true masters are defined by their terms of service – but a review of these written agreements turns up almost as many questions as answers. Second, the law of master and servant, and its subset, agency law, might suggest that users – who chiefly direct these intelligent agents – are their masters, though current agency law’s treatment of software as a “mere tool” of its user is perhaps an oversimplification of this understanding. Finally, competition law and its “single entity” doctrine focus less on form than on function; however, intelligent agents may perform both economic and non-economic functions that might require more varied treatment than competition law provides. As a result, this Article concludes that, as an initial step, focusing on the degree of economic versus non-economic function and interest should be used to determine the standard to be applied to decide the question of for whom intelligent agents work.