With Great Data Comes Great Responsibility. This is the theme for this year’s Data for Good Exchange (D4GX), to be held September 24 at Bloomberg headquarters in New York. And it came about in the way so many good ideas do: in a flash, at a local coffee shop.
Susan Kish, a member of the steering committee for Data for Good Exchange, is now executive-in-residence at Excel Ventures. But when she headed cross-platform initiatives for Bloomberg, the nearby Little Collins café on Lexington between 55th and 56th was her go-to coffee shop. That’s where she and Gideon Mann, Bloomberg’s head of data science, met to think through the big questions and topics for this year’s D4GX. Kish, Mann, and Arnaud Sahuguet — who is the director of The Foundry at Cornell Tech and the third member of the steering committee — had already made some good progress. But after the U.S. Presidential election, there was a surge of concern around fake news and the role of data occasionally leading to false conclusions.
Kish says that led to a conversation about the potential for something akin to a Hippocratic Oath, but for data scientists. Just as medical professionals swear to uphold certain standards, data scientists need to understand the responsibilities inherent in working with data and algorithms. “I still remember Gideon looking up from his notebook and saying, ‘You mean with great data comes great responsibility?'” recalls Kish. “And we said, ‘Oh, that’s it. That’s exactly right.'”
Anyone who works with data, Mann says, needs to understand that algorithms are not neutral. “People often think that if a machine is doing it, it’s fair,” he says. “And often, that’s not true.”
Kish sees an additional and related meaning to the theme, pointing out that great data is not a given. Data is often messy, poorly structured, and full of mistakes. “Getting great data is hard,” she says. “If you are actually given custody of wonderful data, you have a responsibility for that.”
Awareness of these issues, say the steering committee members, has increased substantially since the first Data for Good Exchange, held in 2014. “If you asked me about awareness a year or two ago,” says Mann, “I would have told you that it’s small. But at this point, there is a broad recognition that it’s important.” At this year’s event, he says, there will be papers and panels that deal with real-world applications, where people are grappling with what it means to work with data and be ethical and fair. “That makes me very encouraged,” says Mann.
D4GX provides a unique forum to discuss these ideas. There is no other forum today, says Sahuguet, where people from academia, government, and industry can come together to discuss the use of data to solve social problems. The event, he says, tries to be a new forum where “we can exchange ideas and present research in the spirit of sharing, not in the spirit of selling stuff or getting tenure. That’s what I like about it.”
Sahuguet personally enjoys attending the conference “to learn about problems I have zero knowledge about.” And he encourages faculty and researchers at Cornell Tech to learn about them too. “We can think about how we might apply solutions that are well-known to problems that are totally different.”
The organizers expect that people working in different areas of data science will find the conference useful and enlightening in different ways. Sahuguet thinks that academics may be inspired by new problems to tackle. Those in non-profits and governments might become more aware of state-of-the-art research and the possibilities opened up by data. “They have the problems, and they may have the data, but they may not know how to use one to help solve the other,” he says. Those in industry may find interesting problems to tackle or possibly fund.
Ultimately, all three agree, the event is about building community, about meetings and connections that wouldn’t otherwise happen. Mann says that as people present work and ideas, both formally and informally, he expects them to reach across the aisle and start to develop relationships outside of their normal business. “It’s funny,” he says. “There are papers and panels, but that’s not what it’s about for me. The papers and panels are just a way to get those connections.”