Cedars, April 2019

April 2019 8 OFF CAMPUS “one-third of occasions where these medica- tions were co-prescribed were inconsistent with guidelines.” This is a staggering level of misdiagnosis, and algorithmic sorting of the patients involved while protecting their private information could help prevent such mistakes in the future. These are clinical examples, but with the popularization of smart watches with bi- osensors that can track blood pressure, ox- ygen saturation levels, heart rate and other health factors, medical data is now moving beyond the hospital and into the day-to-day. While the scope of these sensors is limited, the information collected could be used by companies or third-party specialists to help with disease prevention, tracking outbreaks, and alerting users of any potential issues they may develop. Already, Apple is exper- imenting with sending Apple Watch users notifications if their heart rate falls outside a certain range for more than 10 minutes. Some have suggested that this data should be available to insurers as an option for preventive care or to doctors to help de- tect patterns that could factor into a diag- nosis. However, others are concerned about medical privacy and may even fear being discriminated against based on their sensi- tive health history. Data mining has other positive uses outside the medical field, including the realms of sociology, psychology, economics, communications and agricultural research. A 2018 study published in the scientific journal PLOS ONE found that based on soil moisture, temperature, rainfall, pH and or- ganic carbon, data mining algorithms could predict CO2 emissions with more than 80% accuracy, helping farmers effectively protect against climate change. In this light, our data seems a reason- able price to pay for the advancement of so- ciety. Even if it seems worth it, though, it’s still important to recognize what is happen- ing and where it goes — especially because sometimes, companies can mess up. Com- puter science professor Patrick Dudenhofer said this may be a significant risk as more and more companies collect user data. “Unfortunately for most data that’s be- ing stored, it’s not a matter of if it’s going to be hacked, but when,” Dudenhofer said. “And the key feature of cybersecurity is that once the data is gone, it’s gone. Once it’s out there, it will never be private again.” It’s a risk for databases at all levels: internet and technology businesses like T-Mobile, Netflix and Facebook; secure ser- vices like Equifax, Blue Cross Blue Shield and Marriott; government agencies like the Internal Revenue Service, the Office of Personnel Management and even the postal system. These organizations have the kind of data that includes not just your name, address, and phone number, but your credit information, your medical history and your children’s social securi- ty numbers. However, the reality is that users don’t have much of a choice to opt out. Ev- eryone is subject to the IRS — and in fairness, for many secure organizations, hacks are the exception rather than the rule. Other com- panies, though, are less careful. Just this month Facebook acknowledged that it had been storing millions of user passwords in unencrypted plain text, meaning any of thousands of Facebook employees could find them just by searching. Think about how many of your passwords are identical or even just similar to your Facebook pass- word (might want to change that). Yet it’s hardly their first data failure, which is significant because Facebook has become more than just another social media platform. First, it’s become nearly universal among populations with regular internet access. Forbes ranks Facebook fourth on its list of the world’s most valuable brands. Sec- ond, it’s a focal point for personal data, from the information you fill out for your profile to the political articles you comment on to the pictures friends post with your name tagged at a particular location. Third, con- sider how many other online accounts you can sign into using Facebook as a launch- pad. In fact, some companies specifically force users to do this so they get more data than they otherwise would. “It didn’t hurt them in the long run to lose your account data,” Dudenhofer said. “It’s a matter of, do you trust that organiza- tion to both use that information responsi- bly and to keep it secure from other organi- zations? Which historically has been a ‘no’ on both counts.” Most companies assure customers — or the segment of customers who are informed enough to raise concerns — that whatever data they collect is safely anonymized, stored in aggregate chunks detached from individual identifi- ers. This may be true, but as technology has advanced, it has become clear that ano- nymization is not enough. Way back in the mid- 1990’s, graduate student La- tanya Sweeney decided to test the system. She purchased a copy of the voter rolls of the city of Cambridge for $20, which gave her personal de- tails on every voter including name, birth date, sex, address and ZIP code. Comparing the voter rolls to “anonymized” state health re- cords, she was able to send the governor of Massachusetts his entire medical history, in- cluding prescriptions and diagnoses. In 2000, Sweeney went on to demon- strate that only three pieces of information — ZIP code, sex and birth date — were need- ed to uniquely identify 87% of all Americans. As Ars Technica editor Nate Anderson pointed out, “Almost all information can be ‘personal’ when combined with enough oth- er relevant bits of data.” Since then, anonymization has margin- ally improved, and the amount of accessible data has exploded. While optimistic about many possible applications of data science, Gollmer expressed concern that the increas- ing amount of data available to, for exam- ple, employers or insurers may be unfairly detrimental to some people. “It’s going to be there whether you like it or not,” Gollmer said. “But all that infor- mation is not necessarily indicative of who you are right now. When people pull infor- mation so far back, it’s used to give the im- pression that this person was this way back then and he or she is no different today.” The European Union has begun to im- plement legal standards to protect consum- er privacy to a degree, but the United States has yet to show interest in following suit. There’s a tenuous balance between what falls under an person’s “right to be forgot- ten” and what is a matter of public record. Lawmakers are trying to decide where that balance lies, as well as the technical practicality of regulating data collection and use. In a world where technology moves much faster than Congress, who wins the battle over data collection could guide the future power structure. According to Univer- sity of Rochester professor Dr. Adam Frank, the “digital breadcrumbs” created by an per- son’s digital activity are mere particles in the larger picture, denoting the general patterns of society or its specific segments. “Once you see the patterns,” Frank wrote in an article for NPR, “you can understand the world’s behavior. Once you understand behavior you can predict it. Once you can predict behavior you can control it. That is the true promise—and danger —of big data.” Data is a promise because it is criti- cal to the technology our society operates upon on a daily basis. It is a danger because of what happens when used incorrectly. If data is compromised, private information can be exposed, credit cards can be co-opt- ed and identities can be stolen, which is why so many are concerned about its collection, legitimate or otherwise. Data collection is exciting for some, terrifying to others and probably unavoid- able for most. While many expect useful services, public platforms and unlimited entertainment to be provided free of charge online, it’s important to recognize the reali- ty that every user of these sites and services pays a real price — just not a financial one. Breanna Beers is a sophomore molecu- lar and cellular biology major and a staff writer for Cedars. She loves exercising cu- riosity, hiking new trails, and quoting The Princess Bride whether it’s relevant or not. Kassie Kirsch is a junior English major and a staff writer for off-campus news. She loves the outdoors, animals of all kinds and reading more books than she should. Data collection is exciting for some, terrifying to others and probably unavoidable for most.