Jack Dorsey, Twitter’s CEO, is shaking and shaping the online world quite often. However one of his recent attention-grabing steps is more important to our whole society then one might think at first. At the end of October he announced on his Twitter account he’s going to stop all political advertising on his site with a claim: ‘We believe political message reach should be earned, not bought,’ and
His explanation revolved around morality and current political challenges in the online environment: misleading information, disinformation, deep fakes and electoral influence. According to Dorsey, it is not credible to say Twitter is working hard to stop people from spreading disinformation – and then when someone pays for it just let them do whatever they want. This is of course a nudge into Mark Zuckerberg, CEO of Facebook, who just a few days before this announcement confirmed to the American Congress that Facebook is not fact-checking the political advertisements and doesn’t plan to regulate them. Twitter is playing the ‘good guy’ card very well and Google already announced to follow in its footsteps. A great decision and leap forward in the fight of disinformation, one might say. So, why am I not happy about it?
First and foremost, Twitter didn’t announce yet how it is going to define what is political advertisement. And they have a good reason: it’s near impossible job if they really want to stay fair. Of course, politicians and political parties shouldn’t be allowed to pay. Then we have the obvious case of political candidates, right? But when exactly does one become a candidate? Okay, let’s rather focus on content. Everything related to political decisions and topics can’t be sponsored. So NGOs or activists can’t for instance sponsor climate change campaigns, legit. But what about oil companies sponsoring their ‘research’ on how much they are not harming the environment, is that political? And can’t pretty much any topic be politicized? Who will decide? And are the platforms going to coordinate the definitions together, or will it differ from platform to platform, which might cause an imbalance where some are banned everywhere and others just somewhere?
Twitter, Google and whoever will join them is certainly going to present some array and adapt it on the go, maybe even with the help of machine learning (and we’ll get to that in a moment). Regardless of my doubts that it can truly cover everything and thus be fair, I’m much more anxious because of the idea that multinational corporations without a clear accountability are going to define political topics, rather than because of manipulators, who are trying to buy their way into defining them. And hand in hand with that comes my second concern: the artificial intelligence.
Being quite the techno-optimist, I’d be usually seen protecting the development of AI without hesitation. What I’m concerned about is the deployment of AI where the current state of machine learning is not sufficient enough for the task (yet). Take the infamous EU’s copyright directive and its biggest controversy, making the provider of services responsible for uploaded content, therefore making him eligible to implement upload filters – which are yet unable to truly distinguish copyrighted content from, for example, very legal satire or memes. With political advertisement it might become the same story. You feed your algorithm some variety of rules and let it learn. After a while not even its creators will be able to tell precisely why some ads have been blocked and some weren’t. But politics is not a black and white issue. It’s deeply rooted into our society and even though philosophers can discuss the existence and meaning of homo politicus or try to draw a line, current AI is not ready to answer for them.
Last but not least, we need to talk about the principles of disinformation spreading and organic reach. If you want to increase the chance of success of your disinformation campaign, you have to follow the playbook: put into a big bowl of lie a hint of truth, dash of bizzare, some useful idiot to increase the credibility and don’t forget to conceal your hand. And repeat over and over, till you get one of the lies viral. Do you think that the operation infection, the pizza gate, the migration-related hoaxes and whichever big lie you’ve heard in the past years has succeeded because someone paid Facebook, Twitter or Google an outrageous amount of money? Unfortunately, no. You’ve heard these lies because they were skillfully crafted to reach you organically, and no amount of money can compete with that. If I now sponsor counter-tweets to Trump’s tweets in which he is spreading hoaxes, he’d still have bigger and better reach.
In conclusion I don’t believe that ban on political advertisement is capable of treating all the actors fairly, that it’s safer in regard to shaping of the society more than just leaving it be and even effective in its attempt. So, what am I suggesting instead? Several steps. There is one big black hole no CEO wants you to focus on: organic reach algorithms. How exactly is Facebook shaping your bubble? What information is Twitter pushing forward in your feed? What does it mean ‘relevancy’ of content for Google? If these companies really want to take responsibility for their impact on society, they should stop being secretive about their algorithms that are shaping the society, as well as stop to simply shifting the focus towards advertisers. Facebook’s steps towards transparency of advertisement – showing who paid for it and who is being targeted – is ironically a more rational approach than pure bans and it would be my second recommendation. Though Zuckerberg also lacks in any efforts to make the algorithms transparent (where transparency includes explanability, so that we can understand what are the corporations doing and hold them accountable if necessary).
I would like to wrap up my argument with an appeal – to media literacy. The problems we face sound incomprehensibly huge and people are quick to judge and offer short-term solutions. Ban this, regulate that. Politicians consequently failed to focus further than goes their electoral term. That needs to change – in cooperation with schools and NGOs, the governments should start educating their citizens and preparing them for the information age they already live in. Is it late? Yes. Is it too late? Not yet. Hand in hand with well-thought and not blindly regulating legislation and following data, not emotions, we can keep on living and thriving in the information age, without losing the ‘information’ part.