The Wheaton Wire

Joerg Blumtritt, Professor of Practice for Digital Media and Communications, is currently in charge of the Wheaton Institute for Interdisciplinary Humanities (WIIH). The WIIH is a process here at Wheaton where professors can bring ideas for the focus of a discussion series, where they can host lectures, collaborative round table, other themed events, and run courses related to their selected theme. They are given a small budget in order to bring speakers to campus, virtually or physically, and otherwise fund their themed programming.

For his theme, working on the idea with Professor and Department Chair of Film, New Media, and Communications, Patrick Johnson, Blumtritt chose “AI Humanities.” Explaining the concept of AI Humanities, Blumtritt focused on AI uses for creative works rather than automated tasks, and how that  necessitates a conversation about what this use means for humanities fields, drawing on humanities principles including philosophy, ethics, law, and many more.

Blumtritt originally thought the intersection of humanities and AI was something valuable to look into when “7 or 8 years ago [he saw] that there was a rise in interest…and [he] noticed the applications…were much more [focused on] natural language processing, [in particular] understanding or trying to make sense of text.”
Blumtritt provided an example in that“there is a lot of humanities work [being] done with AI and machine learning systems, for instance the deciphering of charred scrolls of the Villa dei Papyri” in southern Italy, near the gulf of Naples (Herculaneum, now known as Ercolano to be exact). These scrolls are “philisophical texts that can be deciphered [using] advanced image processing,” and likely would not be very accessible otherwise. The use of AI in humanities project such as this is critically important to discuss.

“Reclaiming AI for the humanities,” is the concept that AI fundamentally changes the way academic work is done, and that engaging with this idea is inherently a question to be answered by scholars of the humanities. The above project is an example of this, and considering how each field of study involved in such a project is impacted by the use of AI is important. In completing projects, AI and machine learning systems sometimes as “toy models of how systems like the brain might function.” However, AI and humans are not alike.

Perhaps the most accurate way to look at the differences betweeen humans and AI is with language and information acquisition. “You can show [children] one or two examples and they can generalize from that; no neural network will ever do that,” said Blumtritt.

Blumtritt says this raises important questions, primarily“What is the alien-ness that AI has?” The speakers who have visited the WIIH to talk about this topic have insights.

The first talk was given by Vincent Warne, editor of Millenium Film Magazine in New York, who has written “poignant critiques on the most buzzy AI art projects.” Per Blumtritt, Vincent Warne does not naively condemn AI, but draws on the perspective of philosopher Vilém Flusser to adress humanities role in AI.

Flusser wrote about the future of writing and photography in the 1970s and 80s, and was quite good in seeing the role of devices. In his book Does writing have a future? he was quite pessismistic, saying that writing might just go away once machines would be able to do writing in a “sufficiently good way,” framing writing as a code that could be learned by machines.

Of this, Warne raised the question: What does that leave the author to be after that? This is one such question that must be answered by humanities scholars.

The second talk was given by Bruce Sterling, who has twice won the Hugo Award, a literary award science fiction, and was on the cover of the first ever Wired magazine. He spoke about Italian author Primo Levi and his stories about the “versificatore”,  an electronic poet and invented evil corporation in his stories, that could automate poetry when given prompts.

Sterling drew on HB Lovecraft’s literary idea about “the old ones” an ancient race on Earth millions of years before humans who had creatures that served called shoggoths, which were faceless blobs, performing many necessary tasks including underwater construction. At a point in time, the shoggoths mutated and killed their overlords, building their own cities in the ruins.

In a way AI will be like this, a faceless blob but with a human facade, however Sterling was clear AI itself won’t kill people, as it has no will of its own, no joie de vivre.  Rather, it is a machine or tool that will be used and tweaked by others in order to fufill personal goals and wishes.

Lisa Lebduska, Professor of English, Chair of the Department of English and Philosophy, and Director of College Writing at Wheaton also has insightful analysis into understanding AI and its limitations, and how humanists are to play a role. Lebduska is currently teaching a senior seminar in which students are reading fiction that directly engages with issues of social media and AI, and are “running some experiments, doing work, trying out different products, seeing what their limitations are, and then reading about different aspects of AI.”

When asked about what “reclaiming AI for the humanities” means to her, Lebduska said that “the values of the humanities are part of the conversation around AI, whether its the philosophers talking about the ethical implications,” or the call for kill switches on AI because we simply don’t fully comprehend their limits. This raises a question for Lebduska; “Is this almost an existential crisis for humanities?”

Addressing a potential approach if it indeed is a crisis, Lebduska offered that “part of [dealing with AI] will be embracing a kind of nimbleness and fluidity…sometimes I think the humanities are misperceived as being very rigid, [like] we’re holding on to what is and what will always be, and should never change.” Lebduska went on, saying this is not how the humanities is, and just because we may be perceived this way, we should not dismiss AI as something for computer scientists and technologists to deal with. We should take an active role.

Taking an active role herself, Lebduska said that the end-of-course goal for her senior seminar is for her students to “not be afraid of this, and yet [not] be just consuming it mindlessly, thinking about when to use it, [and] what it does for our cognitive development,” so that they, in short, are “practic[ing] critical AI literacy.”

A similar opportunity is being offered next semester, when Blumtritt will be teaching Understanding AI, which will serve as a conceptual introduction to AI, seeking not to stay on the surface but also not to get lost in the math and computer science. He wants to introduce applications of AI that are not “ready made shiny boxes,” as such AI has limits to its use. To explore technologies for creativity, but not just by tweaking prompts, but by maybe coming up with something that doesn’t need prompts.

There will also be a conversation on November 19th, for students interested in the theory, practice, and future of AI, for which more details will be coming soon. With these action items in mind, Lebduska offer a positive note to end on, perhaps a guide as we go through the near future and are interacting with AI.

“We shouldn’t be approaching from a perspective of fear or threat, but rather [in a way that] we have something really useful to contribute.”