Personal data is also used by artificial intelligence researchers to train their automated programs. Every day, users around the globe upload billions of photos, videos, text posts, and audio clips to sites like YouTube, Facebook, Instagram, and Twitter. That media is then fed to machine learning algorithms, so they can learn to “see” what’s in a photograph or automatically determine whether a post violates Facebook’s hate-speech policy. Your selfies are literally making the robots smarter. Congratulations.
The History of Personal Data Collection
Humans have used technological devices to collect and process data about the world for thousands of years. Greek scientists developed the “first computer,” a complex gear system called the Antikythera mechanism, to trace astrological patterns as far back as 150 BC. Two millennia later, in the late 1880s, Herman Hollerith invented the tabulating machine, a punch card device that helped process data from the 1890 United States Census. Hollerith created a company to market his invention that later merged into what is now IBM.
By the 1960s, the US government was using powerful mainframe computers to store and process an enormous amount of data on nearly every American. Corporations also used the machines to analyze sensitive information including consumer purchasing habits. There were no laws dictating what kind of data they could collect. Worries over supercharged surveillance soon emerged, especially after the publication of Vance Packard’s 1964 book, The Naked Society, which argued that technological change was causing the unprecedented erosion of privacy.
The next year, President Lyndon Johnson’s administration proposed merging hundreds of federal databases into one centralized National Data Bank. Congress, concerned about possible surveillance, pushed back and organized a Special Subcommittee on the Invasion of Privacy. Lawmakers worried the data bank, which would “pool statistics on millions of Americans,” could “possibly violate their secret lives,” The New York Times reported at the time. The project was never realized. Instead, Congress passed a series of laws governing the use of personal data, including the Fair Credit Reporting Act in 1970 and the Privacy Act in 1974. The regulations mandated transparency but did nothing to prevent the government and corporations from collecting information in the first place, argues technology historian Margaret O’Mara.
Toward the end of the 1960s, some scholars, including MIT political scientist Ithiel de Sola Pool, predicted that new computer technologies would continue to facilitate even more invasive personal data collection. The reality they envisioned began to take shape in the mid-1990s, when many Americans started using the internet. By the time most everyone was online, though, one of the first privacy battles over digital data brokers had already been fought: In 1990, Lotus Corporation and the credit bureau Equifax teamed up to create Lotus MarketPlace: Households, a CD-ROM marketing product that was advertised to contain names, income ranges, addresses, and other information about more than 120 million Americans. It quickly caused an uproar among privacy advocates on digital forums like Usenet; over 30,000 people contacted Lotus to opt out of the database. It was ultimately canceled before it was even released. But the scandal didn’t stop other companies from creating massive data sets of consumer information in the future.
Several years later, ads began permeating the web. In the beginning, online advertising remained largely anonymous. While you may have seen ads for skiing if you looked up winter sports, websites couldn’t connect you to your real identity. (HotWired.com, the online version of WIRED, was the first website to run a banner ad in 1994, as part of a campaign for AT&T.) Then, in 1999, digital ad giant DoubleClick ignited a privacy scandal when it tried to de-anonymize its ads by merging with the enormous data broker Abacus Direct.
Privacy groups argued that DoubleClick could have used personal information collected by the data broker to target ads based on people’s real names. They petitioned the Federal Trade Commission, arguing that the practice would amount to unlawful tracking. As a result, DoubleClick sold the firm at a loss in 2006, and the Network Advertising Initiative was created, a trade group that developed standards for online advertising, including requiring companies to notify users when their personal data is being collected.
But privacy advocates’ concerns eventually came true. In 2008, Google officially acquired DoubleClick, and in 2016 it revised its privacy policy to permit personally-identifiable web tracking. Before then, Google kept its DoubleClick browsing data separate from personal information it collected from services like Gmail. Today, Google and Facebook can target ads based on your name—exactly what people feared DoubleClick would do two decades ago. And that’s not all: Because most people carry tracking devices in their pockets in the form of smartphones, these companies, and many others, can also follow us wherever we go.