For years, people have warned of the potential for artificial intelligence (AI) to take over activities traditionally considered to be uniquely and quintessentially human. At a time when machine learning and AI are rapidly gaining new capabilities and becoming increasingly present in our daily lives, how close have we come to this future?
In the first Stuttgart "Zukunftsrede," (lecture about the future), organized through a collaboration involving the International Center for Cultural and Technological Studies at the University of Stuttgart (IZKT), the Hospitalhof/Evangelisches Bildungszentrum Stuttgart, and the Literaturhaus Stuttgart, bestselling novelist Daniel Kehlmann (Tyll, Measuring the World) read an essay in which he explored AI's ability to replicate one of humankind's greatest achievements: the ability to create and tell stories.
Following the reading, HLRS Director Michael Resch joined Kehlmann for a nearly hour-long conversation moderated by journalist Eva Wolfangel that focused on how humans perceive and interact with AI, and on the differences between human and machine-based creativity. The event was broadcast online from the Literaturhaus on February 9, 2021.
Kehlmann's essay, titled "Mein Algorithmus und Ich" (Me and My Algorithm), centered on an invitation the author received to travel to California's Silicon Valley to write short fiction in collaboration with an artificial intelligence program called CTRL ("Control"). After Kehlmann entered sentences into the program, CTRL used natural language processing algorithms to respond with sentences intended to continue a story. In an iterative fashion, this dialogue between man and machine generated some surprising and potentially fruitful turns, Kehlmann explained, but ultimately ended in software crashes when the program proved incapable of mastering narrative continuity. Reading some results of this collaboration, the author described his experience of interacting with the machine and what it taught him about the state of AI today.
In the discussion that followed, Kehlmann and Resch engaged in a wide-ranging and lively conversation that complicated the problem of distinguishing between how human artists and computers work. What is the difference, Resch asked provocatively, between a computer that suggests new texts based on statistical analyses of previous texts and a novelist whose writing is grounded on knowledge gained by reading other authors? Speaking from the perspective of a writer, Kehlmann acknowledged that even the most creative artists only occasionally have original ideas, and that human consciousness itself often operates mechanistically.
For Resch, this insight offered a potential way to think about the relationship between human creativity and AI. "We humans only occasionally have these moments when creativity takes place. That is the original meaning of the word intelligentsia. Because of this, it might actually not make sense to talk about artificial intelligence. If instead we talked about artificial logic or artificial rationality, then we wouldn't have a problem, because then we wouldn't be ascribing human intelligence to computers."
What is needed, he suggested, might be a return to the original meaning of intelligence. "On the one hand intelligence has something to do with creativity," he continued, "but on the other simply with the ability to add or multiply numbers. Computers are wonderful at the latter and a lot of what we call intelligence is simply mechanistic, systematic activity. But the other part, what was once called intelligence in the sense of the original Latin expression, that is where we need to ask ourselves, is that what makes us human?"
— Christopher Williams