The privacy paradox: why do people keep using tech firms that abuse their data?

Despite privacy scandals, Facebook is more profitable than ever – journalists must use the tools of tech to understand why

A dark shadow looms over our networked world. It’s called the “privacy paradox”. The main commercial engine of this world involves erosion of, and intrusions upon, our privacy. Whenever researchers, opinion pollsters and other busybodies ask people if they value their privacy, they invariably respond with a resounding “yes”. The paradox arises from the fact that they nevertheless continue to use the services that undermine their beloved privacy.

If you want confirmation, then look no further than Facebook. In privacy-scandal terms, 2018 was an annus horribilis for the company. Yet the resultsshow that by almost every measure that matters to Wall Street, it has had a bumper year. The number of daily active users everywhere is up; average revenue per user is up 19% on last year, while overall revenue for the last quarter of 2018 is 30.4% up on the same quarter in 2017. In privacy terms, the company should be a pariah. At least some of its users must be aware of this. But it apparently makes no difference to their behaviour.

For a long time, people attributed the privacy paradox to the fact that most users of Facebook didn’t actually understand the ways their personal information was being appropriated and used. And maybe that is indeed the case for many of them, for example, new Facebook users in poor countries for whom the Facebook app represents their entry point into the networked world. But that surely cannot be the case for users in western countries. Can it?

A few months ago, journalists on the New York Times embarked on an interesting experiment to see whether internet users really understand the comprehensiveness and granularity of the data-harvesting techniques that underpin surveillance capitalism. Last week, they unveiled the details of the experiment. The reporters picked 16 categories, such as “registered Democrats” or “people trying to lose weight”, and targeted ads at people in those categories. But instead of trying to sell cars or prescription drugs, they used the ads to reveal the invisible information itself. So targeted users would receive ads saying things like this: “This ad thinks that you’re trying to lose weight and still love bakeries. You’re being watched. Are you OK with that?” Or: “This ad thinks you’re female, a registered Democrat and are likely to vote for the sitting president”. Each ad was annotated to indicate the source of each of the inferences it contained.

The point of the experiment, one imagines, was to prompt the question: “How do they know this?” in the target’s mind. The aim, one supposes, was to illuminate the privacy paradox by exploring the cognitive dissonance – the psychological stress experienced by a person who holds two or more contradictory beliefs, ideas or values – the ads triggered in those who received them.

It’s an ingenious stratagem, but unfortunately the NYT doesn’t report whether the ads had this effect. It’s conceivable that the researchers did follow up with the targets and that a subsequent report is forthcoming. But at the moment, we’re left in the dark about the crucial question of whether knowledge really does change people’s behaviour.

Still, it’s a start. We need more experiments such as this, because until we have a comprehensive explanation for the privacy paradox, we will remain in the dark about how best to tackle the menace of targeted advertising. At the moment, the most persuasive explanation of it is what psychologists call “privacy calculus”, the idea that social media users understand the tradeoff between losing privacy and the benefits they get from using services that undermine it and regard the latter as outweighing the former. It seems that demographic variables play a minor role and only gender has been found by researchers to weakly predict privacy behaviour.

The NYT experiment has lessons for journalism too. In an algorithmically curated world, reporters need to tool up if they are to have any hope of holding tech companies to account. And here and there are promising stirrings in the media undergrowth. In this context, ProPublica, a nonprofit organisation based in New York, has been doing great work developing algorithms and bots to explore what Facebook and Amazon are up to.

The time is coming when at least some investigative reporters will have to be not only tech-savvy, but able to use data analytics software as well. For too long, the companies have been able to pull the wool over the eyes of regulators and media organisations that were overawed and intimidated by the complexity of digital capitalism. As the old adage puts it: when the going gets tough, then the tough have to get going.

The Guardian 5 May 2019