Social media’s power demands an advertising rethink
Given that machine learning can take human manipulation from powerful to nearly perfectible, we need strong limits
It’s time for a major rethink of advertising. In fact, an enormous rollback of this runaway industry is in order. It is clear that advances in machine learning and data harvesting have changed the game completely.
But at the same time, knowledge about human limitations has also progressed. We now know how human vulnerabilities and psychological quirks make us susceptible to manipulation.
This is in obvious ways like making us feel inadequate about our bodies, to less obvious ways to do with hijacking primitive reward circuitry in our brains to make us do and buy things that don’t actually make us happier.
In fact, we now even know from work in economics, including my own, that people can become misguided in their explicit beliefs about what makes for a better (happier) life, overall.
Economists have also recently emphasized the idea of a “dark side of capitalism:” just like our capitalist system is remarkably successful at filling niches when and where a need arises, it is equally able to create fake needs whenever there exist human weaknesses that can be exploited. The market is certain to be just as good at that task, which hurts people but makes profit, as it is at the task of helping people while making a profit.
So there is a good basis for being careful about what we allow advertisers to do and to tell people. In Canada, we are among those countries that already have some protections against advertising to children, and stronger online privacy laws than exist in the United States.
Those give us some protection against the frightening intrusions of personal autonomy and privacy for which Facebook CEO Mark Zuckerberg had been defending his firm to the U.S. Congress over two days this week.
But is Zuckerberg’s opinion on regulation really a good starting point?
Given that machine learning can take human manipulation from powerful to nearly perfectible, we need strong limits. In fact, targeting ads to individuals’ profiles should simply become illegal.
Displaying ads that are tailored to a particular product, such as this newspaper, is fine and is how things used to work.
Displaying selective ads on a website based on the search term used to summon that particular website is also not terribly intrusive. But beyond that, the line should be drawn, so that ads may not make use of information from our past, our other online activity or any other behaviour not related to the immediate moment.
This would mean that advertising would become slightly less efficient. Ads would be less perfectly targeted to us, but they would still be effective enough. Advertising revenue might also be spread back out a little more, rather than captured narrowly by whoever has the biggest database and neural network.
What would this mean for Facebook? It would mean the phase-out of the Facebook model. In the future, new micropayment systems may make it more feasible to pay for some of our online services, rather than sell our attention and minds to providers. And other online services may become public infrastructure.
But if targeting ads to individuals was forbidden, there would be no more incentive for the creation of ethical time-bombs like Facebook.
Mark Zuckerberg knows who Facebook’s customers are. They are the advertisers, not its users. And Zuckerberg’s responsibilities are primarily to stockholders. Its profit is maximized when it does whatever is necessary to create a machine with as much power as possible to control the thoughts, beliefs and desires of everyone of us, everywhere.
That sounds like a time-bomb we want to defuse early on.
In Canada, we are among those countries that already have some protections against advertising