�糵����

How do algorithms shape the information we see? What is the impact of human influence on algorithms? The inaugural Lewis College Roundtable explored these questions and more on October 29, 2015. The 90-minute event featured a thought-provoking discussion among a diverse group of panelists—Angela M. Cirucci, Jason Resch (CS ’06), Christian Sandvig, Nick Seaver—who each brought a different area of expertise to the table.

“It’s amazing that computing has evolved so much that we are even having this conversation,” said Sandvig, an associate professor of communication studies and information at the University of Michigan, before the event. “I am interested in examining how computers can be accountable for decision-making. What qualities of computers are making decisions for humans? How are algorithms sorting through cultural data and feeding us information?”

Cirucci added, “Algorithms have become a buzzword recently. I’m interested in understanding how people define them and how they talk about them.” As an assistant professor of communication studies at Kutztown University, she studies the philosophy of social networks and examines how the architectures of those networks shape our identities and personal narratives.

The roundtable conversation, moderated by , dean of Lewis College, opened with a discussion about Facebook, a common example of the use of algorithms. Many people recognize that Facebook filters the information we see, but many are not aware of why or how that happens.

“On Facebook, when you click the like button on a post with a link, you are forever associated with the beginning part of that URL. This data is recorded and used to determine what ads and suggested likes show up on your screen,” Sandvig explained.

“Algorithms change rapidly, and that change is often algorithmic,” Seaver said. “When thinking about Facebook and algorithms, it is actually this enormous, complex thing. At any one time, each of us experiences a slightly different Facebook. Algorithms are used to create different test groups and interfaces, which help Facebook determine what keeps users on the site the longest.”

Seaver, an anthropologist who will join Tufts University as an assistant professor in January 2016, is interested in how algorithms function in human domains, such as music recommendation systems like Pandora. “There is a mismatch between how humans approach the world and how algorithms approach the world. We must recognize the human role in the creation and use of algorithms,” he said.

Resch, a software architect at Cleversafe, creates algorithms that sort data. He also develops them to add security features to database systems. His work has resulted in more than 130 patents issued and 310 patents pending. “Algorithms have limitations. They are not creative or moral, like humans, and they never second guess decisions; they just follow a series of commands,” he responded. “Algorithms can make mistakes, unforeseen by their creators, hundreds or thousands of times in a second without time for human interventions.”

“Whatever is created by humans has human value in it. If there are 150 people working on one algorithm, then the values of those 150 people are embedded into the system. Every human system is non-neutral,” added Cirucci.

Libby Hemphill, assistant professor of communication at Lewis College, attended the roundtable and commented that more transparency is necessary in the digital space. “We don’t need to understand what the algorithm is doing, but why it is doing it. There needs to be more transparency in what goes into making an algorithm—what values were put into it and why?”

The event concluded with the final question: Can algorithms replace the human judgment process?

“Yes, if we consider that the brain is a machine, then we can eventually get there,” Resch responded. “We’ve seen a lot of examples in the news recently, such as driverless cars. As machines begin to eclipse human decision making, it will be interesting to see how computers take over.”

As technology continues to evolve, it is clear that algorithms are here to stay and will continue to play significant roles in our lives.

Group photo (L-R): Nick Seaver, Angela Cirucci, Christine Himes, Jason Resch, and Christian Sandvig.