Thank you.
I think I want to leave you with one message today in my opening remarks, and that is that I really believe that the issue you're diving into of the particularities of the vulnerabilities that were shown and demonstrated through the case of Cambridge Analytica's use of Facebook and collection of data about American and Canadian citizens is not a case of individual bad actors that need to be countered, but rather is a function of structural problems in our very digital infrastructure, which I think are creating weaknesses in our free and open society. These weaknesses, I think, are being exploited by corrupting the quality of information in our public sphere, which is increasingly digital, by magnifying divisions in our society and by undermining our democratic institutions themselves. I want to talk about those problems and the structural elements of these problems by making four points over the next few minutes.
The first is that I think it's really important as a baseline to recognize that there has been a real evolution of our digital infrastructure, particularly of the Internet, over the past 30 years. In very broad sweeps—obviously, this is a much more detailed evolution—the first iteration of the Internet, web 1.0, really did give voice to a whole host of actors and individuals and groups who were excluded from our mainstream public discourse.
Web 2.0, in the 1990s and 2000s, the social web, connected people in really powerful ways and often democratizing ways, as we saw through the Arab Spring and through a whole host of social movements that leveraged these technologies around the world in incredibly positive ways.
I now think that the Internet is something qualitatively different. The problems you're investigating are representative of this difference. I think we're in a third phase of its evolution, what I broadly call the platform era. I would argue that this current version of the Internet is largely controlled by a small number of global platform companies, and for many people in the world the Internet they experience is filtered via these platform companies. That's what I want to talk a little bit about today.
The second broad point I would make is that in this platform ecosystem, this platform Internet, there are two structural problems embedded in that very Internet infrastructure. The first is the way that platforms or the Internet or we have been monetized—what's often called the attention economy or surveillance capitalism.
I would argue that in this tightly controlled market for our attention, audiences can be microtargeted and behaviour can be nudged by anyone from anywhere for any reason. Our attention and our behavioural change is the product being sold in this digital economy.
At the same time as our microtargeting behaviour is being affected or changed, since engagement is the primary metric of value in this attention economy—how much we engage, whether positively or negatively—platform algorithms prioritize entertainment, shock, and radicalization over reliable information. This is embedded in the business model. This is why research shows, for example, that misinformation spreads further and faster than genuine news. It's because it's embedded in the model.
The second structural problem, I think, which we're on the front end of and is going to become a much bigger issue over the coming years, is that the character and what we experience in this digital platform ecosystem is increasingly determined by unaccountable artificial intelligence systems.
These AI systems are used to filter the most engaging content to us, to know what will rile us up and engage us, to determine what we see as an individual user and whether we are seen and heard inside these platforms. Increasingly, AI is used to create versions of reality itself. They're often called deep fakes or synthetic media. A whole new reality is shaped by AI and targeted specifically to us as individuals.
Those are what I see as the structural problems here.
The third point I want to make is that I think these structural problems are responsible for the negative externalities we're now seeing in our democracy, one of which is represented by the Cambridge Analytica case and the 2016 U.S. election, but I think these negative externalities extend far more broadly. Let me describe a few.
One is that the quality of the information we receive, or the information in our digital public sphere, is becoming increasingly unreliable. The platform web is increasingly a toxic place. Highly gendered and racialized speech is incentivized, political discourse has become more extreme and divisive, which you experience intimately, and speech has been weaponized, with a resulting censoring effect. Voices are simply drowned out by abuse. At the same time that this is happening, the digital public sphere is becoming more toxic. We're seeing the increasingly rapid collapse of the industry of journalism, providing weaker and weaker backstops against this flood of false and toxic content.
In my view, democracy requires a grounding of common and generally trustworthy information, and I fear that because of this structural problem this is slipping away from us.
The second negative externality I want to mention is fragmentation. On platforms, we're each given a customized diet of information designed to reinforce and harden our views. The result is that polarization and tribalism can very quickly emerge in this ecosystem. This is a problem for a wide range of reasons, but perhaps most worryingly because it's increasingly leading to actual physical manifestations of individual and collective violence.
A recent study found that in any German town where per-person Facebook use rose to one standard deviation above the national average, a tax on refugees increased by about 50%. I think that Canada without a doubt lags on some of these trends and the social implications of them that we've seen in other western democracies, but fragmentation based on unreliable and microtargeted information is sure to divide us on the issues that are most poignant in Canada now. Imagine climate change, indigenous rights, pipelines and immigration all being fuelled by this structural vulnerability.
The third negative externality, which I think is of acute interest right now in Canada, is the vulnerability of our elections themselves. I would argue that by using the very tools provided by the attention economy, foreign and domestic actors alike can powerfully shape the behaviour of voters. AI and data-driven microtargeting is incredibly powerful during elections, as we saw with the Cambridge Analytica case. Acute cyber-attacks and hacking are a vulnerability, as we saw during the Clinton email leaks or the Macron leaks, but I think you can also be more subtle. I wouldn't want to focus too much on just these very acute public cases.
I can give you an example of a more subtle case. A recent study found that long before the 2016 U.S. election, Russian government-connected accounts created a host of fan pages on Facebook for prominent African-American figures. They did one for Beyoncé and one for Malcolm X. The goal was to build an organic community. They published fan content about Beyoncé to try to build the followers of that page. In the days before the election, they then weaponized that community and pushed content to them designed to suppress the African-American vote.
How do we deal with something like that? How do we even know that this is a foreign-sponsored fan page and that it will be weaponized in the days before the election? This gets at the real structural problems we're facing here.
In the final and fourth point I want to make, I want to offer a few reflections on the public policy solutions to this problem or the governance challenges that this presents.
The first point I would make about public policy here is that it's very clear that self-regulation has proven and will continue to prove insufficient for the nature of this problem. I would argue that the apt analogy is the lead-up to the financial crisis, where the financial incentives are powerfully aligned against meaningful reform of the ecosystem. These are publicly traded and largely unregulated companies whose shareholders demand year-on-year growth.
This growth simply may or may not be aligned with the public interest, and that's how democracies function. When there are negative externalities of largely unregulated monopolies, governments engage to protect the collective good. I think that's where we are now.
I have a second point about public policy here. To me this is primarily a demand-side problem that requires a comprehensive policy approach. Many have argued that it's actually the users' fault, that it's a supply-side problem, that we're consuming and producing toxic content and therefore we should change consumer behaviour. I actually think that misses the structural aspect, and indeed, almost every major global commission or report that has looked at this issue has argued that a comprehensive policy approach is needed. There's not just one silver bullet to this. It's about reforming how we regulate and engage with our digital economy writ large. This is going to involve—