By Eva Hill
hilleva@grinnell.edu
From “Blade Runner” to “Her,” “Black Mirror” to “Made for Love,” artists and scientists alike have long been fascinated by the potential for human beings to form romantic bonds with artificial intelligence (AI). But how would the social and legal complexities of a human/AI relationship play out in the real world?
That’s the question the Grinnell College Debate Union sought to answer in an open debate last Tuesday evening, Feb. 16: should human beings be allowed to marry AI? The Valentine’s Day-inspired event involved two teams of two debaters each, David Dai `23 and Filip Matic `24 for the proposition (in favor of the idea) and Meredith Benjamin `24 and Eleanor Corbin `24 for the opposition (against). (Editor’s note: Eleanor Corbin is a news editor for the S&B)
And while the question may initially seem simple — or unnecessary — the debaters’ questions depicted a much deeper and more complex world of philosophical, moral and legal arguments surrounding human/AI relationships that prod at prescient issues of gender, class, consent, power dynamics and family structures. The debaters noted afterwards that for the audience-oriented debate, the club thought it would be interesting to have gender-separated teams arguing on a subject that plays on many gender dynamics. It’s typical for the group to assign roles in each argument, meaning that debaters’ points are focused more on making the best and most logical argument rather than reflecting their personal views.
“Especially with controversial topics, we make sure it [assignment of roles] is completely random,” Dai said.
Dai opened the debate with a timed seven-minute statement that touched on marriage as a human right for AI that are sufficiently advanced to be able to form human relations, then moved on to the therapeutic and practical potential of human/AI marriages and protection for humanoid AIs under nondiscrimination laws. Corbin countered with a statement of the same length arguing that marriage to an AI would be fraught with consent issues, not least that someone who wants to marry an AI could program it to be unable to reject them. Further, Corbin asked, can AI beings be considered able to willingly consent to anything at all — or can only the programmers who created them do that?
“It gets us right up against both things that we feel really passionate about, things that remain kind of unanswerable even about us, let alone about them,” said Professor Karla Erickson, sociology, on the questions of human rights and free will for AI. Erickson is a feminist ethnographer of labor who is currently working on an upcoming book, “Messy Humans: A Sociology of Human/Machine Relations,” which will explore human interactions with technology and the social effects of AI in human life.
“We don’t really understand where our own sentience comes from, and it makes it very hard for us to assess the value of another kind of sentience,” Erickson said. “Especially when we don’t know where ours comes from.”
Legal issues of considering AI as human beings are extremely complicated when considering marriage as a human ritual, Benjamin argued in a later portion of the debate. Some qualities of AI could make any human/AI marriage vulnerable to annulment: If multiple people marry AI based on the same program, is that bigamy? If a person marries an AI based on technology that is only a few years old, even if the AI technically has the intelligence of an adult, is that child marriage?
Meanwhile, Matic argued that with the assumption of an AI so advanced a human could fall in love with it, the program would need to allow it to withdraw consent, ask for a divorce, and practice other abilities to reject requests similar to humans. It’s a point that illustrates an issue that Erickson describes as “almost tautological”: if AI are equivalent to humans, they should have human rights, and if they have human rights, they must be able to reject other humans — so if a person who wanted to marry an AI was doing so because they didn’t want to marry a human, what would happen if and when the AI rejected them too?
The debating union will continue to hold open debates on campus in the coming weeks, as well as attend an off-campus tournament (possibly traveling to New York this semester for one such competition).
“We’re going to try to make them [the open debates] as relevant as possible,” Benjamin about picking topics to debate. “It’s actually easier to pick ideas when we have some sort of parameter.”
The Grinnell College Debate Union meets at 7 p.m. on Tuesdays and Thursdays each week. Students interested in joining the club can email debate@grinnell.edu or talk to Professor Mark Baechtel, anthropology, who advises the student debaters.
