Thomas Hobbes on Artificial Intelligence
Personhood, and Responsibility: A Hobbesian Perspective
Artificial Intelligence, Personhood, and Responsibility: A Hobbesian Perspective
by Beezone
The following summary is drawn from Chapter XVI (“Of Persons, Authors, and Things Personated”) of Thomas Hobbes’s Leviathan. Although written in 1651, Hobbes’s distinction between the “Actor” and the “Author,” and between natural and artificial persons, remains notably and strikingly relevant to current discussions concerning the governance of artificial intelligence, agency, and legal responsibility.
The interpretation presented below — while in no way assuming legal authority — is broadly consistent with current legal principles and precedents governing artificial intelligence as presently understood. Under current law, AI systems are generally not recognized as independent legal persons possessing sovereign agency, moral accountability, or autonomous authority. Responsibility continues to reside primarily with the human beings and institutions that create, authorize, deploy, interpret, or act upon AI-generated output.
In Chapter XVI of Leviathan, Hobbes distinguishes between a “natural person” and an “artificial person.” A natural person speaks and acts on his own behalf, while an artificial person represents the words and actions of another through authorization. Hobbes further distinguishes between the “Actor,” who performs or speaks, and the “Author,” who owns and assumes responsibility for the action.
Applied to artificial intelligence, this distinction suggests that AI systems may function as artificial or representational agents, but they are not true “Authors” because they possess no independent consciousness, will, intention, or moral responsibility. AI systems generate language, but they do not own, intend, or stand behind what they produce. Responsibility therefore remains with the human beings or institutions that create, authorize, deploy, interpret, or act upon AI-generated output.
Under Hobbes’s framework, authority and accountability belong ultimately to the human “Author,” not to the artificial mechanism that merely produces or represents speech. Unless society someday grants AI independent legal personhood, AI remains an instrument of mediated human agency rather than a fully responsible person in its own right.
In conclusion, artificial intelligence — or alternative intelligence, machine learning, machine intelligence, expert systems, thinking machines, algorithmic intelligence, or whatever term one chooses to use — can never properly be said to constitute an independent “person,” nor in many instances even a true “agent,” regardless of condition, case, history, present circumstance, or future development. Its computational output does not originate from sovereign moral consciousness, but from systems designed, trained, authorized, prompted, interpreted, and operationalized by human beings. Therefore, following Thomas Hobbes’ Intelligence, both the input into such systems and the consequences of their output remain, initially and finally, the responsibility of the real human person who operates, deploys, delegates, or acts upon the technology.