Consumer A — once seen by marketing firms as a generic soccer mom — is now known, thanks to an analysis of her online activity, as a 35-year-old who jogs two times a week and prefers building blocks over dolls for her two elementary-aged daughters.

Digital dossiers compiled from the bits and bytes of Web searches, online purchases, fitness trackers and other Internet-connected tools and services are transforming a once-distant relationship between consumers and corporations into a more direct interaction.

But what happens when the tools that help marketing firms steer Consumer A toward sales on running shoes use her purchasing history to decide which shoes she can afford to buy? What if her health insurance company decides a biweekly run isn’t enough activity to offset her bad eating habits and demands a third day in exchange for lower premiums?

An examination of the data economy’s potential for abuse is the crux of a long-term research project that will be conducted by Carnegie Mellon University professor Alessandro Acquisti during his two-year fellowship through the Carnegie Corporation of New York.

Advertise with us

Mr. Acquisti, a professor of information technology and public policy, co-director of the CMU Center for Behavioral Decision Research and renowned data privacy expert, last month was selected as one of 32 members of the inaugural class of Andrew Carnegie Fellows.

The fellows, who will research topics ranging from policing and race to the safety of generic drugs, are following the inaugural theme of “Current and Future Challenges to U.S. Democracy and International Order.” Fellows receive up to $200,000 each toward their studies. A total of $6.4 million was awarded this year.

For Mr. Acquisti, who has conducted award-winning research surrounding data analysis, privacy and social media since arriving at CMU in 2003, studying the potential for discrimination within data analysis is a natural transition from his earlier work.

An October paper written with CMU Professor Christina Fong discussed the potential for hiring discrimination through social networking sites, while a 2011 paper demonstrated how “facial recognition, data mining algorithms and statistical re-identification techniques” can give savvy cyberthieves access to critical information such as Social Security numbers.

More critical analysis of how these digital dossiers can be used against consumers is essential during a time when corporations and consumers alike generally tout the benefits, said Mr. Acquisti.

Advertise with us

“The reason I’m interested in the research is there seems to be a moment of faith in so-called big data as a solution for societal problems and something that’s uniquely a force of good, an economic win-win for all parties involved,” he said.

“While I do believe that more data available about behavior and becoming better at analyzing the data can certainly help us improve services and products — and ideally address, if not solve, certain societal problems — I also expect the companies that use data could end up creating changes in allocation of surplus.”

Laura Brandimarte, who coauthored a paper discussing the potential for abuse with Mr. Acquisti, agreed that extensive study into this particular downside of data collection is overdue.

“If the market for information is not carefully regulated, big data can lead to a serious imbalance of power between individuals, whose information can be so easily exploited for profit, at times even unbeknownst to individuals, and companies, organizations and governments that have the upper hand,” she said in an email statement.

The plan, Mr. Acquisti said, is to use scale modeling, data studies and experiments in targeted advertising to understand which scenarios will lead up to the win-win outcome and which could encourage consumer abuse.

Advertise with us

The notion that social media data can be used to practice social discrimination was confirmed in the October study, which revealed that job candidates identifying as Muslim on social media were less likely to receive calls for interviews from employers based in areas with a majority Republican voting base.

On the consumer side, the idea of using data analysis for price gouging is one that immediately came to mind as a potential for trouble.

“The economic lingo goes: once a seller knows exactly how much you want a good, once a seller discovers your so-called reservation price — the maximum you’re willing to pay for a good — the seller will likely charge you exactly the price, which means they’re trying to extract surplus from the interaction with you,” Mr. Acquisti said.

Even without illegal price gouging, Mr. Acquisti sees potential for creating a system of “winners and losers” by using data from fitness monitors and Internet-connected cars to set health and auto insurance premiums.

He took that idea a step further with discussion of startup companies being acquired by tech behemoths such as Google and Facebook planning to combine the new data with existing stockpiles.

Advertise with us

Pointing to black-box driving programs such as Nationwide Insurance’s SmartRide, which awards discounts to drivers who install devices to collect information on speed, distances driven and other statistics, Mr. Acquisiti said the phenomenon is already underway.

What he’s most concerned about is a near-future when emerging technology companies with no explicit plans to share data change their policies.

“Maybe from an economic perspective you could consider this efficient, but there is this little and very important issue of consumers not being aware at all that this is happening and not being able to foresee the consequences of buying a fitness bracelet,” he said.

Mr. Acquisti said he’ll kick off work for the fellowship sometime later this year, but he expects his research will go far beyond the two-year threshold. If his hunches are correct, he might be still uncovering new findings as the hypothesis underlying his research begins to unfold en masse.

“I would be surprised if it doesn’t happen in the next five years,” he said.

Deborah M. Todd: dtodd@ppost-gazette.com or 412-263-1652. Twitter: @deborahtodd