By Eric Sorensen, WSU News
PULLMAN, Wash. – The Internal Revenue Service is breaking several laws by mining large data sets and combing through social media posts in its search for people to audit, a Washington State University professor says.
– Informed when the government is collecting data on them.
– Given the chance to review and correct the information.
Such policies are required by the fair information practices incorporated into the Privacy Act of 1974.
“They’re not doing that at all, and they’re certainly not informing people when they’re doing it,” said Houser. She makes her case in a 55-page paper in the summer issue of the Vanderbilt Journal of Entertainment and Technology Law.
Texts and social media
According to information obtained by the American Civil Liberties Union, the IRS also has violated the Electronic Communications Privacy Act and legal precedent by obtaining electronic communications without a warrant. This practice, authorized in the IRS audit manual, contradicts the 2010 U.S. v. Warshak ruling, which reaffirmed citizens have a reasonable expectation of privacy in their emails, and the government needs a warrant to obtain them. The IRS agreed in a Senate hearing to cease reviewing emails but said nothing about texts and social media.
Additionally, the Computer Matching and Privacy Protection Act requires that each agency “maintain in its records only such information about an individual as is relevant and necessary to accomplish a purpose of the agency” and to verify the accuracy of the data it uses.
Chances are high that the purchased data and mined data amassed by the IRS are going beyond this limit, said Houser. Moreover, the data might be incorrect, said Houser, as people often “airbrush” what they say about themselves and post incorrect information about others.
That’s particularly true when people post information on line, said Houser, author of the 2013 book, Legal Guide to Social Media: Rights and Risks for Businesses and Entrepreneurs.
IRS quietly using algorithms
Traditionally, the IRS looked for tax returns to audit by targeting math mistakes or mismatches with third-party documents, like W-2 wage and salary forms. But after budget cuts prompted the creation of the Office of Compliance Analytics in 2011, the service has relied more heavily on data, including commercial and public data pools and social media sites like Facebook, Instagram and Twitter. Meanwhile, the IRS says little about the Office of Compliance Analytics and how it works.
“Taxpayers should be aware that what they say and do online could be used against them in IRS enforcement actions,” said Houser. “If the IRS is using data mining on Facebook or other internet sites to locate potential noncompliant activity, this would violate, at a minimum, the federal consent requirement.
Houser is concerned about potential discrimination that can arise when the service uses algorithms to seek perceived irregularities in tax returns. Some groups of people might have spending patterns and deductions that get flagged for auditing when in fact they have done nothing wrong. While audits are supposed to be random, algorithms based on faulty data could lead the IRS to repeat previous instances in which it appeared to use audits for political purposes. Most recently, the IRS was accused of targeting conservative organizations affiliated with the Tea Party.
- Kim Houser, clinical assistant professor of accounting, WSU Carson College of Business, 509-335-7385, firstname.lastname@example.org
- Eric Sorensen, science writer, WSU News, 509-335-4846, email@example.com