“If you’ve ever met a human, you know humans don’t act rationally like computers,” philosopher and theologian LeRon Shults told a UCSB audience last week.
By Melody Li
At a time when technology and data are impacting society to a degree not seen before, Norwegian theologian LeRon Shults argues that the humanities have a crucial role to play in the future applications of artificial intelligence, and the ethics that must accompany that.
“If you’ve ever met a human, you know humans don’t act rationally like computers. We’re biased, anxious, and we change,” said Shults last week at a lecture hosted by UC Santa Barbara’s religious studies department and the UCSB library.
Shults is an author and professor of theology and philosophy at the University of Agder in Kristiansand, Norway. He is part of an international team of computer scientists, philosophers, and religion scholars who work on an interdisciplinary project of human simulation.
Traditional AI allows scientists to predict the future based on a computer model, but human simulation means the predictions can include our biases and hopes.
The researchers use AI to create artificial societies filled with thousands of virtual people, called “agents,” who mimic human behavior based on cognitive psychology. Researchers then program these agents to have different group identities and to mimic the attributes and beliefs of a real country’s population.
“Our project tells you what happens in the real world when you have these conditions,” said Shults.
The goal of prediction models is to give politicians a forecasting tool in a simulated society before trying out policies in the real world. With the push of a button, world leaders can see how policies would actually affect society and make the best decisions based on these simulations.
Norwegian theology and philosophy professor LeRon Shults, at UC Santa Barbara last week.
“Every model is a simulation of human ethics. We can see the causes and consequences of unethical behaviors,” said Shults.
Shults conceded human simulation can be viewed as a dangerous idea in light of work done by groups like Cambridge Analytica, a consulting firm that used data to help the Trump campaign as well as the Brexit movement in Britain in 2016.
But he believes that his group’s research is both more powerful than that of Cambridge Analytica and at the same time is more explicit in purpose.
“Our goal is to make all our models transparent and put simulation platforms online so anybody can use it,” said Shults.
His team was able to predict whether and how religious beliefs and church attendance in certain European countries and Japan would change over a 10-year period from 1998 to 2008 by studying 1998 data from those countries.
“Our models were up to three times more accurate in prediction than linear regression analysis, which is the next closest competitor,” said Shults.
The project, Forecasting Religiosity and Existential Security with an Agent-Based Model, examined the conditions that lead to a more religious or a more atheist society.
“Our model shows that educational homophily, hanging out with people who are more educated or differently educated, is one of the key mechanisms that speeds up the secularization process,” said Shults.
The project relies heavily on the expertise and collaboration of humanists and scientists. Shults strongly believes that in order for the humanities to survive in a culture where STEM disciplines increasingly dominate, it is important for the humanities to collaborate with the technological sciences to solve real world problems.
“I would suggest that the humanities connect with these disciplines and provide something they can’t provide, which is a deep understanding of human nature, of the processes of history, and of the very dynamics of society,” said Shults.
Melody Li is a third-year Communication major at UC Santa Barbara. She is a Web and Social Media Intern for the Division of Humanities and Fine Arts.