Industry Experts Talk Potential Regulation, Legal Issues of Algorithms

Algorithms normally put out information or conclusions that aren't "super surprising," University of Maryland College of Information Studies associate professor Jennifer Golbeck said, and shouldn't be relied on as the ultimate source for a decision. "The things we have to keep in mind with algorithms today is that they are going to tell us stuff, but we absolutely have to have intelligent humans taking that as one piece of input that they use to make decisions," Golbeck said. "[And] not just handing control over to the algorithms and [letting] them make decisions on their own because they are going to be wrong a lot of the time. They are not going to do things as well as a human can do." Ian Bogost, Georgia Institute of Technology professor of interactive computing, urged the media to delve deeper into the actual processes behind the operation of algorithms, rather than simply equating them to an all-knowing, mystical being. "The way we discuss algorithms in the media really does matter," he said. Laura Moy, New America's Open Technology Institute's senior policy counsel, said when thinking of the problematic outcomes of the innovative uses of algorithms, "a lot of times there aren't really clear legal 'don'ts,'" and pointed to consumer privacy as another issue presented by algorithms and data collection. Moy said algorithms have the potential to perpetuate human biases and could have "some sort of disparate impact" on users, also saying it's difficult to identify or correct the addition of human biases in algorithms. It would be worth thinking about building in ways we can check for bias and to identify it when using algorithms to produce a service or result, she said. "At a basic level, transparency about what information is going in and how it might be used to make decisions that could impact the individual, that level of transparency to the individual is important." Moy also said regulators are looking into correcting bias in algorithms in their design, and a few federal agencies have been thinking about it as a fairness issue and starting to address it. "From the regulators' perspective, full transparency, full insight into what all of the inputs [into algorithms] might possibly be and [in]to how it works is important," said Moy.