Resurrection or distortion? How AI can preserve – or annihilate – Holocaust history
As UNESCO issues warning on potential for AI to change the historical record, author Todd Presner probes ethical concerns behind ongoing merging of AI tech with humanities research
Reporter at The Times of Israel
Amy Kurzweil never met her Holocaust survivor grandfather, but the graphic memoirist recently conducted several dialogues with him through an AI-based chatbot.
“Could I get to know him? Could I even come to love him, even though our lifespans didn’t overlap?” said Kurzweil during a TED Talk about the project.
By feeding 600 pages of her grandfather’s letters, lectures, and other writings into a computer, Kurzweil built a “selective” chatbot. In other words, the dynamic bot was designed to respond to questions using information from her grandfather’s writings.
“If we married AI with my grandfather’s writing, we could build a chatbot that writes in my grandfather’s voice,” said Kurzweil.
As a young man, Fredric Kurzweil was able to flee Austria because an American woman at one of his piano performances offered to sponsor his visa. After becoming a prominent music educator in the US, Kurzweil died in 1970, aged 57.
“I wanted to talk to my grandfather because he, like me, was an artist,” said Kurzweil, the daughter of iconic futurist Ray Kurzweil.
Kurzweil has published two graphic memoirs based on her family, including an account of her grandmother in the Holocaust and a project in which Kurzweil collaborated with her mother.
“I think AI has a special role to play in the mission of memory,” said Kurzweil. “[AI can help] ward off annihilation by animating the legacies of our families and our cultures,” Kurzweil told TED participants.
“People don’t just disappear when they die. AI swirls our conception of time and space. It can remix and extend our identities,” said Kurzweil. As with cartooning, AI can “help us appreciate the vastness of humanity, if we let it,” she said.
While AI expands the boundaries of Holocaust memory, however, nefarious forces are deploying the same technology to erode the memory of the genocide in which six million Jews were murdered during World War II.
In June, UNESCO published a paper titled, “AI and the Holocaust: Rewriting history? The impact of Artificial Intelligence on Understanding the Holocaust.” In the paper, researchers identified several problematic AI technologies that pose a risk to Holocaust history.
“The threats associated with AI on safeguarding the record of the Holocaust are manifold, including the potential for manipulation by malicious actors, the introduction of falsehoods or dissemination of biased information, and the gradual erosion of public trust in authentic records,” said the UNESCO paper.
In his recent book, “Ethics of the Algorithm: Digital Humanities and Holocaust Memory,” author Todd Presner called for creating best practices to address some of the threats identified by UNESCO.
“I’m not a pessimist or alarmist with regard to these new technologies, but I do think that humanists — including historians, ethicists, and scholars of culture and society — need to help shape them and help shape how they are used for research and education,” Presner told The Times of Israel.
A professor of European languages at UCLA, Presner said AI is blending with traditional research methods to form a new-old discipline called digital humanities. His book helps visualize the merging of AI with Holocaust research. For example, a set of data-based visualizations depicts an analysis of common vocal tones and cadences used by survivors in recorded testimonies.
“Ethics of the Algorithm” places humans at the center of both the problem and the solution. As creators of the codes — or algorithms — behind AI, it’s up to humans to embed ethical considerations into new technologies, said Presner.
“How we receive testimony, in what ways it is mediated, whether we understand it, how we access and interpret it, whether we compare it, what tools we use to analyze it, how technologies and people generate and transform it, how we value and attach meaning to it — these are all issues that bring questions of methods together with questions of ethics,” wrote Presner.
Written for an academic audience yet accessible, “Ethics of the Algorithm” contains dozens of colorful visuals depicting research conducted by Presner and his graduate students. Each illustration has a QR code for readers to examine data sets or learn more about the AI-based experiments.
Presner clearly shares Kurzweil’s enthusiasm for deploying AI to deepen Holocaust memory. At the same time, he wants algorithm creators and other digital practitioners to set ethical guidelines, as advised by UNESCO.
“AI, when used responsibly, has the potential to deepen understanding, create new knowledge, and allow us to process data at a scale previously unimaginable. The issue for me remains authenticity, accuracy, and digital provenance, or where the data and algorithms come from,” said Presner.
