A new survey from Pew Research has given us some insight into how Americans are feeling about our algorithm-driven society, and the verdict is, we aren’t loving it. One question in particular shows that social media companies could help the public feel more comfortable if they would just lose the political campaign ads altogether.
Pew asked 4,594 Americans about their attitudes toward a range of uses for computer algorithms, and it’s clear that respondents have a lot of concerns. Broadly speaking, the survey found that people are worried that algorithms will be incapable of applying fairness if used for “decision-making in situations with important real-world consequences.” Examples in which the majority of respondents found it unacceptable to use an algorithm included criminal risk assessment for people up for parole, automated resume screening for job applicants, automated video analysis of job interviews, personal finance scores using many types of consumer data. And what has more real-world consequences than our elections?
When Pew focused its questions on social media, it found that, depending on a person’s age, between 31 and 40 percent of respondents believe it’s acceptable for companies to use personal data to algorithmically recommend messages from political campaigns. The demographic that found the technique the least acceptable was over 65-years-old. It might not be surprising that younger groups were more comfortable with an algorithm deciding what political propaganda would be most effective in plying their vote, but the youngest group, 18-29, was the second most skeptical with only 35 percent saying the technique is acceptable.
Political campaign advertising has been a hot-button issue for social media since the 2016 elections turned into a convoluted clusterfuck of allegations about foreign interference and messaging to Americans. While major companies like Facebook, Twitter, and Google have implemented policies identifying a “political ad,” we’ve also seen those policies result in legitimate journalism being labeled as if it were propaganda. In Facebook’s case, we’ve also seen its initial method for identifying who is running a political ad completely collapse because it was practically based on the honor system.
Out of all of Pew’s questions about acceptable uses of personal data to make recommendations, political campaign ads ranked at the bottom. Social media networks and algorithm-driven businesses like Google should take note. Much of the failures and complexities of this issue have come from the question: What is a political ad? Some people feel that a sponsored post from a publication shouldn’t be labeled as a political ad simply because its subject matter falls into the category of politics. But identifying content from a political campaign would be more simple. Of course, the big money in politics these days comes from “independent” political action committees, so if companies decided to outright ban political campaign ads, they would have to ban PACs as well.
Yes, it would still be hard to implement, as Facebook’s struggle to properly identify political ads illustrates. And people would almost certainly find ways to game the system. But strong penalties for users who are found breaking the rules could give some level of deterrence, and greatly limiting political ads would go a long way toward removing the most potentially consequential problematic ads on social media. It’s a positive sign that Americans seem to be showing signs of a basic awareness that algorithmic targeting is flawed and dangerous. It might be at its most dangerous when it works well.
In June, Brad Parscale, President Trump’s former digital director and current 2020 campaign manager, told 60 Minutes that 80 percent of the 2016 campaign’s digital budget went to Facebook. Parscale said the campaign used “hundreds of thousands” of different micro-targeted ads to reach voters on the issues they specifically cared about. He said on a typical day they generated 50 thousand to 60 thousand ads. Let’s just say a candidate, hypothetically, doesn’t have a great respect for the truth. With micro-targeted ads, they could hold every position under the sun but only claim a certain position when getting the message to someone who agrees with it. John might see that Governor Patrick is a social liberal with anti-abortion sympathies, and Sally might see that Governor Patrick is a moderate with hard pro-choice views.
Banning political campaign ads could help tech companies stem the growing tide of controversies and make the public feel better about using the platforms. The problem is there’s a lot of money in those ads. Research group Kantar Media estimates that advertisers dropped $5.25 billion on ads for the 2018 midterm elections. Digital ads made up $950 million worth of that spend, but tech companies have a bright future in the market. The 2018 spend was quadruple what advertisers spent on digital ads during the 2014 campaign. Digital measurement firm eMarketer puts Facebook and Google’s share of the total digital market at around 57 percent, and Kantar Media said the two platforms were the preferred choice for politicians. The big up-and-comer in the digital ad space is reportedly Amazon, and it’s unclear if political promotions will ever be up its alley. So the power to steer this out-of-control campaign bus is really in the hands of just a few.