Friday 8 November 2019

Free speech; Mark Zuckerberg; and online 'misinformation'

I've been intending to write a post on free speech for quite some time. I've made several attempts, but it never quite crystallised in the manner I had hoped (I find that if I do not write a post start to finish in a single sitting, it never quite seems to get completed; my curiosity moves on to other matters, my thinking evolves, and when I return to the issue my preferred means of expression changes and necessitates a re-write). There was so much I wanted to say, but I struggled to find a way to present all of my arguments in well-structured manner that was not overly contorted or lengthy.

However, Facebook CEO Mark Zuckerberg recently gave a truly excellent speech at Georgetown University, which eloquently outlined a lot of the core principles I fundamentally believe in that underpin a lot of the analysis I wished to present. I was highly impressed, and recommend people view it. He has made my job a lot easier, as he has already summarised a lot of the foundational principles in an an eloquent and concise manner, so I thought I would merely content myself to add just a few supplementary thoughts in this post. As I noted in a recent Tweet, the world is lucky to have such an enlightened moderate at the helm of such a powerful enterprise.

In recent years, the West has seen an unprecedented (in modern times) cultural assault on the principles and values of free speech. The zeitgeist at present is that the principles of free speech are mostly being espoused by the 'alt right' to justify hate speech and the spread of misinformation, and this perspective has been amplified by the election of Donald Trump, which is widely regarded in liberal circles to have been primarily enabled by the spread of misinformation online/via social media, including via 'Russian trolls'. Many believe the spread of online misinformation is allowing the rise of 'fascism'.

The reception to Zuckerberg's speech - particularly from the US political left - has been remarkably hostile (but not surprisingly so in light of this cultural shift), given that it is basically an espousal of traditional Western liberal democratic values. Core to the accusations Zuck is facing is that he is not doing enough to counter online 'misinformation', and particularly with respect to his stated refusal to censor political ads, or statements made by politicians containing misinformation. Democrat Kamala Harris, for instance, has been calling for Trump's Twitter account to be banned.

There are many causal factors underlying this opposition. One is simply the common human tendency to place short term pragmatism/expediency ahead of more universal principle (something I originally blogged about here). It may well be the case, for instance, that in the short term, promoting free speech does in fact help Trump get re-elected. If you think avoiding Trump's re-election is more important than anything else, you might be willing to say the core principles of Western liberalism, such as freedom of expression, ought to be curtailed, because the ends justify the means. It suffices to say that the 'ends justify the means' approach has paved the way towards (and perpetuated) tyranny many times throughout history. The Founding Fathers were extremely wise to enshrine free speech protection in the constitution, knowing all too well that special interest groups would repeatedly seek to curtail free expression in the name of expediency when it happened to suit their interests.

However, the more fundamental point I wish to make in this post is that a core component of these criticisms of Zuck tie in very closely with a lot of the analysis I presented in yesterday's post on the human tendency towards (and the dangers associated with) binary, black and white thinking. The attacks Zuck has endured all rest on an implicit belief that the world is simple and black and white, rather than nuanced, complex, counter-intuitive, and often contradictory. If you believe the world is simple and black and white, you will also believe that there is a clear and obvious dividing line between what is 'misinformation' and what is not. And if it is obvious something is misinformation, then the case for disallowing its dissemination is much stronger.

The problem is that in the real world, often what is misinformation is not so clear cut, and it can be extremely difficult to distinguish between misinformation and merely differences in opinion, or opinions that are merely unorthodox/non-consensus. And as Zuck points out, granting centralised authorities the power to decide what is misinformation and what is not has historically had the effect of denying people a voice, and reinforcing existing power structures - something liberals and progressives putatively oppose (railing against patriarchy, white privilege, etc). The fundamental issue is: who gets to decide what is misinformation, and how are you so sure that that power will always be exercised in a manner you approve of?

When analysing things from a principled perspective, it always helps to consider how it would look if the shoe was on the other foot. If your principles are sound, they should apply irrespective of which way the partisan or expediency chips fall with respect to any particular issue - it should work in both directions. If it works only in one direction, one is prioritising short term expediency over universal principle. Consider the issue of climate change, for instance (which I also discussed in yesterday's post), and imagine a different world with different social power structures. Imagine a world where Trump supporters were in control of social media platforms, rather than Valley liberals; AOC had just been elected president; and calls for the restriction of the spread of online misinformation were coming not from liberals, but from conservatives.

You could imagine these commentators criticising Zuckerberg for not doing more to counter the online spread of misinformation around climate, which had enabled someone like AOC to get elected. Extinction Rebellion protests were disrupting economies, and the spread of misinformation about how we were in a climate emergency and the world was going to end in 12 years if we didn't take radical action to reduce carbon emissions to zero, were a major threat to the health and prosperity of the global economy, and global poverty reduction efforts.

If Zuck gave the exact same speech, he would be criticised for not taking down these memes. Furthermore, he would be criticised for political ads, and allowing extremists left-wingers like AOC to spread misinformation about a coming climate emergency in order to win votes. Such misinformation ought to be banned. Free speech has its limits, people would argue. If enough misinformation spreads online, really bad outcomes can befall the world, such as misguided efforts to completely ban fossil fuels and carbon emissions, so its is absolutely necessary to take radical action. The ends justify the means, and pragmatism need to trump principle.

See the problem? Who exactly should be tasked with deciding what is misinformation and what is not? And how would liberals feel if social media platforms were controlled by Trump supporters, and they blocked AOC's Twitter account due her being deemed to be spreading misinformation about climate? They would view that as tantamount to fascism, and yet that is precisely what they are demanding Zuckerberg do to Trump in the name of countering fascism. The truth is, what passes as 'fact' and 'misinformation' is often simply a matter of disagreement - you think your ideas are factual, and the people who disagree with you are spreading misinformation. And how are you so sure you're right?

Centralising the power to decide what views are acceptable speech has many attendant problems. For a start, it paves the way for malevolent actors to control the dialogue/media/narrative to entrench their own political power or personal interests, or promote their own political views. But it doesn't need to be that bad. Even if centralised authorities/social media platforms do not end up abusing their power for personal political gain, what will most likely end up happening will be a 'tyranny of the majority', where orthodox views are considered information, and unorthodox views (which are often misunderstood) are considered misinformation. As someone who routinely finds error in orthodox, consensus opinion in many fields, and whose intellectual development has benefited greatly from the absorption of diverse insights from people with a range of opinions and political persuasions, this troubles me deeply.

If we proceed down this path, we will end up creating an echo-chamber of conformity, where innovative thought becomes impossible. You have seen this happen in the Islamic world. Orthodox religious views held by the majority are considered fact, and unorthodox dissident views are considered misinformation (or 'heresy'). It suffices to say that this is not the best path towards a liberal, free, and pluralistic society. People need to be free to hold and express opinions - even inaccurate ones. To be sure, Western liberal democratic values have their problems. It's by no means a perfect system. But we need to be very careful that we do not rush to throw the baby out with the bathwater.

The basic problem here is that all human beings have biases. Human beings do not perceive objective reality, but merely their imperfect interpretation of reality. Reality is far too complex for any one individual to fully comprehend, and so what human beings do is use 'mental models', or simplifying heuristics, to filter reality, and reduce the level of complexity down to a manageable level.

Mental models act as a lens or prism through which people view the world and make judgements, and how somebody interprets inputs/events, and the associated conclusions they draw, will to a large extent reflect whatever assemblage of mental models they have amassed throughout their life (as well as their personal interests). Given the same set of inputs, people that reach similar conclusions on issues generally have closely overlapping mental models (and/or interests), while those who disagree generally have very different mental models, and hence 'see' very different very things.

Mental models per se are not the problem - indeed they are indispensable in helping us function in the world and make decisions. However, the problem emerges when human beings forget that they are viewing the world through a prism, and are not seeing objective reality - merely their interpretation of reality. When you believe you are seeing objective reality rather than merely your own error-prone interpretation of reality, it is all too easy to conclude that people who disagree with you are stupid or have malevolent motives, rather than seeking to understand why they see things differently than you.

That is not to say that some mental models are not better than others, nor that there is no objective reality. Properly applied, the scientific method is a relatively effective mechanism for separating fact from fiction, by demanding beliefs be tested, and be capable of making repeatable, accurate predictions. The problem is that our understanding of reality is still incomplete, and the conclusions of science continue to evolve as we discover new insights. Our mental models are incomplete maps, not the terrain, which omit details and have errors and missing sections. And without freedom of speech, or with an insistence that we already have complete and accurate maps, it is impossible for us to improve these maps over time, and better understand the terrain.

While individuals are subject to biases, the same true to an even greater extent of groups. Indeed, psychology, biology, history and contemporary observation all suggests that groups tend to amplify bias and promote the emergence of groupthink, rather than moderate it (unless what is being described as a 'group' is actually set of decentralized individual nodes). There are many reasons for this. Human beings are 'social learners', which means a large amount of learning takes place through the observation and imitation of other human beings, and a lot of this happens subconsciously, through the uncritical absorption of norms, customs, values, and beliefs.

This can be seen in the fact that more than 99% of people who are religious are the same religion as their parents/communities. If people were rational, independent learners, they would research all the religions and then choose independently which one they wished to adhere to (if any), but people seldom behave in that manner in practice. Genuinely independent thinkers are vanishingly rare.

Indeed, some observers - such as Noah Harari (author of Sapiens) - have argued that homo sapiens ability to engage in mass collective delusions was a critical survival advantage over other, more rational homo species (such as the Neanderthals, who had larger brains), because it allowed large numbers of individuals to coordinate their behaviour, and behave in individually irrational ways for the collective benefit of the group. The more cerebral and rational Neanderthals were no match for raving hoards of irrational, fanatical human beings. Homo sapiens seem to have had the optimal mix of rationality (creative problem solving) and irrationality (the ability to co-ordinate in large numbers and do things that are individually irrational for the perceived good of collective - for instance risking one's life in a political uprising, or charging into battle on the front lines).

Because we are social learners that take refuge in consensus opinion and societal norms, widely held beliefs will therefore tend to reinforce and amplify the biases present in other individual members of the tribe, rather than moderate them, which can lead to a sort of runaway bias inflation, or 'narrative bubble'. This is fundamentally where the much-lamented 'madness of crowds' comes from, in my submission. More moderate biases present in individuals can be amplified and allowed to grow unchecked in a group context, leading to a runaway bubble that can be completely perplexing to outside observers looking on from a safe distance in either space or time. This dynamic explains not only financial bubbles, but many other things as well, from suicide cults and mass political ideologies/movements, to contemporary outrage/cancel culture on liberal US university campuses.

In the Soviet Union, they really thought that people who subscribed to capitalist ideas were mentally deranged. They sent them to correctional institutes for re-education. It seems bizarre to us now, but this is what happens when you get a runaway narrative bubble, and biases amplified by groupthink. What is consensus opinion becomes mistaken for fact, while dissident views are considered misinformation. This is exactly why censoring free speech to remove perceived 'misinformation' is so dangerous. It is how repression/tyranny is born.

In the past, I have expressed concern on Twitter about the trend towards social media censorship. I do not think this reflects express political motives on the part of social media executives. I think it reflects genuine bias and groupthink within their 'trust and safety' departments, which causes them to mistake their political beliefs/opinions/mental models for fact, and confuse contrary opinions with misinformation. I am relieved to see that Zuckerberg has realised the potential danger associated with this trend, and his influence might result in this concerning trend being nipped in the bud. It has taken some courage on his part to stand up to the overwhelming opposition he has faced within the hotbed of Californian liberalism, and his courage and wisdom should be applauded.

The only things which keep humanity's tendency towards irrational groupthink and runaway narrative bubbles in check are external constraints and the forces of decentralisation, and free speech is an important component of the latter (as are individual rights and political autonomy/diffusion). The Soviet Union (and Communist China) both eventually collapsed as their ideological bubbles were in such discord with reality that the gap between the ideology and reality eventually become too large to deny. They eventually reached the point of radical disillusionment, which is necessary for the collapse of narrative bubbles. Ideologies are tenacious, but they do have a breaking point - the point where the level of visible disconfirmatory evidence simply becomes too overwhelming. However, that can take a very long time, and it can require a lot of unnecessary suffering in the interim, and in the absence of external constraints, in a highly centralised society they can persist almost indefinitely.

One of the key safeguards against this tendency is decentralization. All individuals and groups have their own biases, but due to some combination of chance and insight, some individuals and groups will have mental models more closely aligned with reality than others, and those individuals and groups will tend to prosper more than other groups (there are of course other contributing factors as well, such as genetics; superior access to resources, etc). This allows better ideas to percolate from the bottom up over time through natural, evolutionary forces. Centralization thwarts this process. This is why private sector entities are usually far more efficient and adaptable than government and quasi-government entities (and monopoly enterprises are also less efficient and adaptable than those subject to competition). Private enterprises (excluding monopolies) are subject to market discipline; monopolies, and government and quasi-government entities are not.

Nation states are by definition centralized, but one of the protections against the dangers of centralization and political monopoly has been competing individual nation states (or individual states with sufficient political autonomy within a political union, such as in the US). Core to the collapse of communism in both China and the USSR was the comparative US example. The US grew much richer and more powerful, and the gap eventually became so large that it was undeniable that its economic system was superior. It helped hasten the point of radical disillusionment within these societies, and catalysed change from the outside. If the world had only one country, and it became the USSR, Mao's China, or Kim's North Korea, there would be no such outside, decentralized catalyst for change.

In addition, one (of many) of the keys to the US's success has also been the devolution of a lot of autonomy down to the individual 50 states. This allows different states to implement different policies. Some states will do better than others, and people and business will migrate to those states that do better. Poorer performing states then come under pressure to up their game and reform. It creates some necessary competitive tension, which is a necessary counterbalance to the formation of narrative bubbles, which in the absence of competitive tension will always arise over time. Political centralization and monopoly are the enemy of these dynamics, which is also one reason why I oppose moves towards more centralized, global government (in this respect, I think projects like the EU are a step in the wrong direction, as policy diversity is essential to progress).

Another important example of decentralization is capitalism itself, as opposed to central planning. Different businesses are set up by different people/groups with different biases, but some will be less biased than others (through chance or insight), and therefore make better decisions. Those businesses will grow, and subsume the resources previously controlled by weaker businesses, which will fold. Other businesses will also then seek to emulate businesses that are succeeding. Over time, a more rational society better aligned with reality emerges, and prosperity increases.

Freedom of speech is the decentralization of ideas. Without doubt, it has its problems and imperfections. But it has stood the test of time. It is worth remembering that the spread of misinformation and propaganda is not something new - it has been with humanity since the dawn of mass communication. When the printing press was first invented and commercialised, elites and the Catholic Church opposed the spread of books as they feared the broader populous could be easily mislead by bad, heretical ideas and the spread of propaganda.

What actually happened was that books and the democratication and decentralization of information ended up liberating Europe from the dark/middle ages, freeing Europe from yoke of Catholic oppression, ushering in the reformation, the enlightenment and age of reason, the French Revolution, and ultimately the industrial and information revolutions that have lead to modernity. That is a formidable track record. Those that oppose the decentralization of information and democratization of ideas would do well to ponder that fact.