© 2022 WUGA | University of Georgia
shupe_2.jpg
Play Live Radio
Next Up:
0:00
0:00
Available On Air Stations

Athens News Matters: Algorithms vs. People- Who to Trust?

robotTerryNEws.jpg
Terry News
/

You’ve probably already interacted with an algorithm today. They suggest faster routes as you drive to work, recommend songs on Spotify or Apple Music, and in my case help me to procrastinate finishing this article, by watching TikToks.

We may use algorithms all the time, but do we actually trust them?

That’s the question that Aaron Schecter, Eric Bogert, and Rick Watson are all trying to answer

“The problem is generally called algorithmic advice. So it is pretty much what it sounds like in that, people are getting recommendations about anything from an algorithm. And it’s somewhat of an open question as to how people evaluate those sorts of things.”

They ran a pretty simple experiment. Subjects were shown an image, filled with objects. They were then asked to estimate how many objects they could see. After their guess, they were told either an algorithm or a group of people guessed differently.

“People were saying okay I think there are 3,000 objects here. And then it said okay an algorithm says there’s 5,100. And they’d say hmmm, ‘I’ll say 5,000’.”

The goal of the experiment was to see if algorithms, or a group of people were more convincing to the average person

“We saw a much bigger change when it was an algorithm giving the recommendation. So if a human said that you might say I’ll revise upward a little bit, but not by as much.”

So for things like estimating crowds and choosing which stock to buy, people tend to trust an algorithm.

“And again there are things that you know, models should be good at. They take data, they take input, they provide output. I think what we’re trying to show is that people need to be aware of the fact that even if they are giving you junk, you still may have a tendency to believe it. Because it comes from an algorithm”

And that’s the concern for a lot of people. Joe Silva hosts Athens 441.

“We take a modern super eclectic look at modern music. We cover everything from americana to techno.”

Joe’s whole job is to curate playlists and put together a radio show and he doesn’t like using algorithms.

“I’m always suspicious of algorithms because if you use a platform like spotify or something it's always going to drive you to the most common denominator. So its is like ‘oh you like 80s music and here is an 80s playlist, and you might get one thing in a playlist of 100 tracks that you don’t know”

Joe is always trying to find new and interesting music, a process that as of yet, algorithms are not really designed to do. The same goes for Janet Geddis. She is the owner of Avid Bookshop in Athens, which means a lot of her job is recommending books to people.

“More often than not however people come in able to describe the sort of reading experience they want to have, and I and my colleagues at Avid Bookshop are all experts at books. So we are able to narrow down some options for customers based on what they’ve said they want to read… Now that we have been closed to the public for over a year we at Avid Bookshop have had to reinvent how we do those personalized recommendations.”

Avid has added to their website, to try and recreate an in store experience.They also created Avid Bookmatch. It is a questionnaire that customers can fill out, and a selection of books will be hand chosen for them. But Avid Bookshop isn’t using an algorithm to do that.

“Everything we do is driven by humans who are truly empathetic. So I often talk about how anybody I hire to work at Avid loves humans as much as the love books. So we really care about connecting the real person to a book that will provide them the experience they are really looking for.“

Both Joe and Janet like the process of one on one connection, and finding the right piece of media for each person. That’s something an algorithm can’t always do, and that means it’s the next step of Schecter, Watson, and Bogert’s research. People may trust an algorithm to count, but they don’t always trust them to pick an album.