Reinventing liberal democracy in the age of intelligent machines: a talk

Transcript of my talk at Conway Hall Ethical Society of London, and London Futurists.

Let me start with the word “democracy”. If you think that you are citizens living in a democracy you are mistaken. You, I, are citizens in a constitutional republic – or monarchy, as is the case in the United Kingdom and several other European countries. The constitutional republics, or monarchies, that we live in have a form of government called “liberal democracy”. It is the word “democracy” that confuses us. Let’s say, for now, that the word is meaningful only on specific dates in an electoral cycle, once every four years or so, when we are allowed to exercise our right to hold free elections and decide who will rule us until the next election. Between elections we do not have a “democracy”. Why? Because the word “democracy” means the rule of the people, and clearly it is not the people that rule but our representatives, the government we put in place with our vote.

In his book “The Social Contract” Jean-Jacques Rousseau summarizes this important point thus: “The people of England regard itself as free: but it is grossly mistaken; it is free only during the election of members of parliament. As soon as they are elected, slavery overtakes it and is nothing”.

I profoundly agree with Rousseau’s observation albeit without his dramatic overtones. Nevertheless, let me also say that I am a huge fan of constitutional republics and liberal democracies and, for the rest of my talk I will try to defend them to the best of my ability and suggest ways to improve them.

In defence of constitutional republics: rights vs democracy

But before doing so let me explain why I feel that way. It is all about rights. Liberalism, the philosophy that came out of the European Enlightenment during the 17th and 18th centuries, invented the concept of natural rights, which means rights that each and every one of us has by virtue of being born a human. This was quite revolutionary at the time. It is also an idea in direct conflict with democracy. You see in a real democracy, where the people rule their society by majority voting, the many can vote to revoke certain rights from the minority. We saw this clearly in the Brexit referendum, where the 52% decided to diminish the rights of the 48% (I mean their rights as EU citizens, such as to freely travel and work across the EU, appeal to the European Court of Justice, etc.). In very much the same way, in a real democracy, an impoverished majority may vote to confiscate the property of the wealthy few, or take away rights from minorities whose color or religion are different from the majority’s, or from those who are not tall enough or healthy enough, or do not exercise often. You see where I am going with this. Totalitarian regimes and authoritarians are all about real democracy, people rule and all that, there is a lot of excitement in the beginning, flags flying and crowds cheering, until things go quickly south, and one ends up in a concentration camp, or a gulag, or a killing field, or a poverty line.

And that is why I have decided to defend constitutional republics: they protect human rights through the rule of law, parliaments, courts of justice, checks and balances and a free press; but they do so at the expense of real democracy. This tradeoff suggests that, if we want to preserve our rights, freedoms and liberties, the demos, the deplorables, us, must be content with existing in suspended political animation for the duration of the electoral cycle, and be awakened just long enough to confirm in a polling station the rule of the elected few, before going back to sleep.

This existential, and paradoxical, tension between rights and democracy is at the heart of the persisting political crisis that we have been witnessing since the Great Recession of 2008-09.

The deficit of trust

Publics are no longer trusting constitutional republics governed as liberal democracies.  As a result, publics are susceptible to the sirens of populism, which means to the message of politicians, or political parties or movements, that prioritize the will of majorities over the rights of minorities. So publics give their vote to those who will ultimately take away their rights. We have seen this phenomenon before, in the early 20th century. The evils of Nazism, fascism and communism, and the ghosts of their tens of millions of victims, are a cautionary tale of what can go wrong when public discontent is ignored.

But exactly what is it that we are not trusting about our political system?

I would argue that there are at least two principal areas of public mistrust, and fear: firstly, mistrust in the institutions and workings of liberal democracy. This mistrust is fueled by AI algorithms that polarize public opinion, as well as AI systems used by private organizations and governments to spy on our private lives.

Secondly, there is genuine fear for the future, especially because of the impact of automation technologies such as robots, big data and AI, on work and livelihoods.

Let me therefore explore those two areas in turn, examine how intelligent machines are implicated, and suggest ways to repurpose technology in order to evolve -– or upgrade, if you prefer – our system of government.

The principal-agent problem

Political representation is the key decision-making mechanism of liberal democracy and it suffers from the so-called principal agent problem. We the citizens are the principals and the politicians that we vote for are our agents, in the sense that we trust them with running the country on our behalf and in our name. The problem is that the politicians’ interests do not always align with ours. Politicians in a liberal democracy are a professional class that need to keep their jobs; i.e. to get re-elected. One might say that all they need to do is represent us well, do our bidding, look after our interests, and we’ll happily reward them with our vote come next election. Reality, however, has a way to clash with political theory, and good intentions. In the case of liberal democracies reality is influenced by money. What I mean by this is that powerful minorities with special interests and deep pockets can exert disproportional influence on the political process and policy outcomes, such as legislation, regulation, international trade, foreign policy, etc. Simply put, money buys power, and politicians, through lobbying and quite often through straightforward bribery.

By way of hard evidence, a study by Princeton and Northwestern Universities shifted through nearly 1,800 US policies enacted between 1981 and 2002, and compared them to the expressed preferences of average Americans (50th percentile of income), affluent Americans (90th percentile), and large special interests groups. The researchers concluded that the US is dominated by its economic elite. As stated in the research report: “When a majority of citizens disagrees with economic elites and/or economic interests, they generally lose. Moreover, because of the strong status quo bias built into the US political system, even when fairly large majorities of Americans favour political change, the generally do not get it”.

That’s the swamp Trump referred to, so successfully, in 2016 when he won the Presidency. In any case, for large swathes of people in western democracies, evidence and data are unnecessary for what they already instinctively know, that the political game is rigged and stacked firmly against them. That elites – a word meaning anyone except them – are conspiring while lining their pockets in the process. Hence, the mistrust. Enter now, AI algorithms in social media, that act as self-confirming megaphones for biased opinions, or downright crazy conspiracy theories, and what you get is the death of facts and the obliteration of listening. But we need facts, and we need to listen to each other, especially when we disagree. We need facts and civil dialogue desperately, if liberal democracies are to survive in the 21st century. So, what can we do?

Citizen Assemblies

My suggestion is that we need an injection of real democracy in liberal democracies, in the form of consultative citizen assemblies.

Here’s how this works. You pick a group of citizens by sortition. The composition of the group should reflect the demographics of the wider society to which the group belongs, in terms of age, gender, race, religion, etc. Then, this representative group of citizens – the assembly – is given a subject to deliberate and deliver an opinion. For example, a constitutional citizen assembly was set up in Ireland in 2016 to deliberate on changing the abortion law. Over a period of five weekends 99 complete strangers, an assemblage of students, housewives and truck drivers, with views ranging from pro-lifers to pro-choicers and undeciders, paved the way to the national referendum that took place in 2018. The citizen’s deliberations were broadcast on the Internet. They discussed abortion with legal, ethics and medical experts, and listened to testimonies by women who faced crisis pregnancies and had to get an abortion outside Ireland. At the end of the deliberation process the majority of citizens recommended to amend the constitution. Indeed, two thirds of the assembly suggested legalizing abortion without any restrictions.  The view and decision of the citizen assembly reflected the national vote in the Irish referendum of 2018. The Irish citizen assembly thus helped to achieve broad national consensus on a highly polarizing subject.

Citizen assemblies are already being used across the world with great success and in many countries. They have helped in taking difficult decisions on a number of issues ranging from urban planning and health, to managing the environment, public spending and institutional set-up, as the case in Ireland.

I would like to see citizen assemblies becoming a new liberal institution that works in tandem with parliaments and the executive. By including citizens in the formation of legislation and policy, we can dispel much of the mistrust that currently plagues our system of government. I often wonder how much better liberal democracies would have dealt with the Covid-19 pandemic if citizens were also consulted on what measures should be taken, and were included in the deliberation of the possible outcomes of the various strategies and options. The top-down imposition of lockdowns and the curtail of civil liberties have polarized public opinion, and have fueled the spread of misinformation and conspiracy theories. As a result, a recent survey by Pew revealed that 49% of Americans would rather not vaccinate themselves against Covid.

You may of course argue that public health is something for the experts to decide; the epidemiologists, the medical doctors, the statisticians, the “scientists”. And that non-experts, who are most of us, should just sit quiet and trust the experts to guide us out of this mess. But that is not the case. Science can tell you what could be the outcome if you decided A instead of B. But deciding between A and B is always a political choice. Politics are indispensable in human affairs and instrumental in resolving the constant dilemmas that we are faced with. Clearly, we can no longer leave politics to the professional politicians alone. Citizen assemblies are a step towards making liberal democracies more transparent, inclusive, trustworthy, and – yes – more democratic too.

However, if we are to scale up the use of Citizen Assemblies and embed them as a new institution of government at local, national, and international level, we need technology to make the sortition process simple and the work of the citizen assembly easy, especially when citizens deliberate on complex issues that require specialist knowledge. Artificial Intelligence systems, such a new generation of conversational agents, can help bridge the knowledge gap between experts and non-experts, and inform the deliberations of citizens on any subject, however, complex it may be. In my book Cyber Republic I give a detailed description of how this new kind of AI would work in a Citizen Assembly, but let me summarize here the key points. Today, the best algorithms we have are at the service of global advertising colossuses. As such, their goal is to personalize content around our biases and vices.  Hence the negative implications of AI in fueling political polarization. But we could repurpose AI and develop a different set of algorithms that help us understand an issue, check the validity of information before pushing it for our consumption, expose us to opposing views, sensitize us to the various decision costs, and help us develop a new level of civic consciousness, one that compels us to prioritize the well-being of the whole, and encourage us towards consensus.

Fears about the future of work

Let me now switch to the second reason that publics mistrust liberal democracies; which is fear about their economic future and prosperity. This fear is strongly linked to the future of work, and how the combined forces of free markets and automation technologies are eroding the balance of power between capital and labor, and replace labor –white and blue collar workers alike – with intelligent machines. The so-called Fourth Industrial Revolution is upon us, and many people – in fact most people – worry that robots may take away their jobs, and the jobs of their children and grandchildren.

Work has been the bedrock of social and economic mobility since the first industrial revolution, some two hundred years ago. It is through work that globalization managed to raise the living standards of billions in the late 20th century and dramatically reduce poverty across our planet. Work has greatly benefited from technology. Since the first industrial revolution technologies, such as steam engines, electricity, trains and airplanes, computers and telecommunication networks, have uncreased human productivity and created the phenomenon of economic growth – something that was inconceivable in all the previous eras of human history. Machines have extracted humanity from the so-called Malthusian trap where humans living in an agrarian society are bound by the limits of their low productivity.

Machines are physical and cognitive multipliers that augment our productivity and enable us to create surplus that can support a larger human population, which then can produce more surplus, and so on. Machines are the key to transitioning from an economy of scarcity to an economy of abundance. Machines are also disruptors of traditional business models, and some technologies – called general-purpose technologies – can dramatically change the whole of the economy, by creating new industries while destroying others in a relatively short time period.  In 1901 there were 3.25 million horses in England, plowing fields, hauling wagons and carriages, pulling boats on the canals, toiling in the pits, and carrying armies into battle. The advent and democratization of cars in the following decades changed the centuries-old use of horses completely; gone were the professions that supported an economy of horses and in came jobs such as petrol station manager, auto mechanic and car dealer, and of course petrol stations, car manufacturers, the list goes on. Economists describe two complementary effects in times of major technological disruption: the displacement effect (jobs eliminated) and the compensation effect (jobs created). So why should it be any different now, with AI and robots? Shouldn’t we expect the same two effects to come into play, and have new jobs created in the future, better jobs, paying better money?

Of course, there will be such new jobs – we are already seeing the creation of high-value roles, such as data scientists and AI specialists; but I think the better question we must ask, is how many of those new jobs will be created, and how evenly distributed will be in the population. In other words, how will the enormous economic bounty of the Fourth Industrial Revolution will be shared, and if there will be a stake in the emerging AI and Data economy for us all?

The evidence so far has been that most of the wealth created by technology and globalization has gone to the hands of a tiny minority. Moreover, we are currently witnessing the platforming of work and the rise of the gig economy across all professions. The future of work may not be the complete elimination of employment due to automation, but it is very likely to be a future where work is uncertain, intermittent, and not well-paid.

A dystopian view of the future and the Fourth Industrial Revolution would be massive unemployment and underemployment, together with ever-increasing income inequality. We witnessed a prelude of this dystopia during the Great Recession of 2008-2009, and the story is likely repeat itself after the end of the pandemic and the lifting of government subsidies and furlough schemes. It is extremely doubtful that liberal democracies will survive the repercussions of a broken social contract. So what can do about it?

Rethinking work and UBI

Maybe we should rethink what work means for society; rethink work not as the main source of the necessary income to support our living, but as a source of creativity and happiness.

One idea to transition towards this more utopian future is that of Universal Basic Income, whereby a minimum income is given to every citizen regardless of any other income they may have. UBI – as Universal Basic Income is called for short – can reduce the anxiety of uncertain work and provide a cushion to absorb troughs between jobs or contracts. If meaningful, UBI could potentially free humanity from the toil and sweat of working for a living, and usher us into a new age of economic abundance, a new human civilization.

But where will the money come from, I hear many of you asking. Let’s examine the only real example we currently have in the world of a UBI which is the Alaskan Permanent Fund that pays an annual amount between $1000 and $2000 to every permanent resident in Alaska. This money comes from the oil and gas that Alaska has. So, if you have a common valuable resource then you can find the money to finance UBI. If you do not, then UBI needs to be funded through the only other two means that a government has to raise money, taxes and borrowing. In a future without work, where the vast majority of citizens do not have any significant income to tax, the government will have to tax big corporations and try to balance the books through borrowing. Which means that big corporations will have even greater influence over government – since they are footing the tax bill – and citizens are getting deeper into debt. For me this scenario where the government is virtually – or literarily -privatized is indistinguishable from feudalism. Citizens will become the surfs of the big corporates who will be paying us pitons through a meagre UBI, just enough for  us to get by. I do not think such a scenario is sustainable. Sooner or later it will lead to revolution. The yellow vests will be back.

So it seems to me that the only way we can fund UBI and sustain our liberal system of government is to find a valuable resource – or asset – that belongs to all the people, and use that asset to fund a meaningful UBI. What could that asset be?

Data Property Rights and Data Trusts

Let’s quickly look how the global economy has been transforming over the past few decades and focus on the nature of assets, and in particular the difference between tangible assets (think oil& gas, like in Alaska) versus intangible assets (think knowledge, intellectual property, data and algorithms).

You can see clearly the trend from tangible to intangible assets. The AI economy of the 21st century will be based on data, our data. So what if we transformed our collective data into a national, or a regional, or a city asset. Imagine, for example, data from sensors in a smart city, collected by an organization that represents, and is governed by, the citizens of that city and their elected representatives. Call this organization a Data Trust.

The Data Trust can then give access to that data to any private organization that desires use of that data in order to develop algorithms, products and services, for a fee. Those fees are collected and are distributed to the citizens as dividends, as UBI. Data are indeed the new oil, but, unlike oil ,they are an inexhaustible resource whose value increases exponentially when combined with other data. Take for example a data set of patient data and combine it with a genetics data set; the combined value if far greater from the separate values of each set, because it is their combination that can allow for greater insights into human health. And just to put a dollar value in all that, recently American Airlines put up as collateral in order to tap to a US government loan their loyalty program; a dataset that was valued at $30 billion. Given that there are 67 million members in that program (about the same as population of the UK), if that data set was administered by a citizen Data Trust, each member would have a share of $450,000. If we assume a 2% net annual interest on this asset, each member would receive an income of around $9,000 per year. And that’s just the valuation of one data set. A smart city Data Trust could hold dozens of citizen data sets, such as health records, transportation habits, street pollution levels, etc.

By monetizing our data through democratically-governed Data Trusts citizens in liberal democracies can retain their political power, become key stakeholders in the AI economy, and reduce economic inequality without the need of coercive wealth redistribution.

History at crossroads

Ladies and Gentlemen, today I tried to explain why I am at the same time concerned as well as hopeful about the future of liberal democracy. Concerned because of the perfect storm that is set upon us, from increased work automation, enormous economic inequality, public mistrust, fear about the future, and political polarization that so readily explodes into violence. Hopeful because there are ways to rethink data and AI technologies so that we introduce more direct citizen participation into policy-making through citizen assemblies and upgrade democracy, as well as monetize our collective data to fund a more creative and happy life.

History is at a crossroads; I think we all feel this way. The economic crisis of 2008-09 set in motion forces that are changing our world in dramatic ways, at national as well as international level. Those forces are eroding alliances that have kept peace and security since World War 2, unleash revisionist forces in Asia, Europe and the Middle East, challenge our liberal system of government, awaken dormant voices from the extremes of the political spectrum that call for a greater role for the state while ignoring the perils to freedom that come from such a slippery road towards authoritarianism and oppression.

But there is hope. The Covid-19 pandemic has accelerated the transition into the AI economy, which includes the delivery of positive social and environmental outcomes and the rethinking of corporate governance. I personally would like to also see the institution of data property rights and the repurposing of AI so that its goals align with the goals of humanity, so it becomes less autonomous and more integrated into human society. The technological means that we have are out disposal in order to create a better future are enormous, and so is their potential for positive change.

But will that future happen? The fact that technologies are available for us to build a better future does not in itself guarantee a better future. The single, most important driver of societal change is ideology: the web of dominant ideas that a society refers to when thinking through its challenges and seeking solutions. As we face a new era in geopolitics and in the economy, we must move beyond the 20th century dipoles of Left versus Right, and Conservatives versus Progressives. We need to explore radical new ideas and create new political syntheses. Rage – so common nowadays – must be replaced by reason. Tribalism must give way to a new sense of patriotism, and by that I mean the feeling of belonging to a larger societal group that shares a land, a history and a set of traditions, a group that includes people with whom we may disagree, that look different from us, worship different gods or none, and yet we choose, because of our shared patriotism, to prioritize the benefit of everyone above our own. As citizens we must believe that knowledge is possible, that we can know, however imperfectly, and that to know is better than not to know. This is very important to highlight, as truth, facts, and knowledge are under a constant barrage of algorithmic and philosophical denial. Democracy needs objective knowledge and falsifiable truths in order to function and survive, and that is what distinguishes it from totalitarianism that thrives on myths, lies, and philosophical relativism.

We need philosophers, economists, political scientists, engineers, mathematicians, historians, social scientists, and every citizen, to contribute in developing the new politics for the age of intelligent machines, a new set of ideas for rebalancing the relationship between the individual and society; for thinking in terms of networks rather than units; of collaboration as well as competition; of the economic value of social capital and not only of the capital in the bank; of digital platforms whose business models distribute dividends to those who add value through their data, participation and interactions – rather than today’s models that exploit our data and personal networks to create and funnel enormous profits only for the few; of free markets where the incentive to take risk is only strengthened by the incentive to do good.

Thank you very much for listening.

Citizen Assemblies: a TEDx talk

Do you trust your politicians? The people who represent you, and take decisions on your behalf about vital things, like health, education, national defence, and importantly where your taxes are spent Research data show that trust in politicians is at an all-time low across most liberal democracies. In fact, only 1 in 5 citizens trusts their politicians.

But why is that so? Well, first of all, let me say that the 4 out of 5 who do not trust politicians (and I count myself in that group) are right. There is a strong misalignment of goals between us citizens and our representatives, something that is often called the “principal-agent” problem.

In politics we are the principals and our representatives are our agents. They are supposed to do their best to maximize our well-being. But agents have their own priorities: the most important of which is to get re-elected. This means that they must often fall in line with powerful interest groups that influence the outcome of elections. Oftentimes agents, our politicians, must decide which one of their many constituencies to serve first. Various constituencies may have conflicting interests. Take for example what happened during and after the Financial Crisis of 2008-2009. Politicians in Europe and the US prioritized saving the constituency of bankers banks at the expense of the constituency of taxpayers, a decision that in many ways set off the wave of populism we witness today.

So is there something we can do to solve the principal-agent problem in politics, and reclaim trust in democratic institutions? Could we the citizens – the principals – have more of a direct say into policy decisions? Referenda is certainly one way to do so, but they suffer from two very serious problems.

The first one is the asymmetry of knowledge. Not everyone is equally knowledgeable about everything. That’s why we have experts. We need to listen and trust their advice. In an ideal democracy citizens have access to knowledge they can trust, from where they can inform their opinions. But we all know that this is not really happening. Very often the necessary knowledge to understand the complexity of a problem is quite difficult to acquire. And, equally important, can you trust the source? The phenomenon of AI algorithms peddling “fake news” in social media makes this problem ever more acute. In fact it is nowadays so acute that many citizens do not trust the experts. There are people, for example, who contrary to all scientific evidence believe that inoculation causes autism, that the planet is not really warming up, and in extreme cases that the earth is flat and astronauts never landed on the moon.

The second problem is time. Getting involved in politics takes time. Time that needs to be taken out of the little time we have to spend with our loved ones or do the things we enjoy, or indeed do work and earn a living. Learning all you need to learn about climate change, or inoculations, or genetics, or AI, in order to have an informed opinion takes a lot of time. Given the lack of time to participate we will generally opt to form opinions based on the most digestible information available out there, which may not be right. Not having time to participate in politics was the reason we delegated the responsibility of government to our agents, the professional class of politicians.

So lack of time and the asymmetry of knowledge make citizen participation problematic. So we are stuck. Our agents, our representatives, the politicians, will always serve the interests of themselves, their powerful friends and their families above our own. If we do not like them, the best we can do is vote them out every four years. But four years in an interconnected world of accelerated change can be a very long time.

No wonder so many citizens are trusting democracy less and less.

But what if we re-imagined democratic politics by solving those two problems, the knowledge asymmetry and lack of time? There is indeed a way to do so, and it is called a Citizen Assembly.

Here’s how it works. You select a group of citizens by lot, by sortition. This group should be diverse enough to reflect the demographics of the wider social group. Then you bring those people together, and you give them the opportunity to learn, debate and query facts on whatever is that you want the Citizen Assembly to opine on. And you compensate them for their time.

I was personally involved in a CItizen Assembly made up of 140 citizens from 9 different European countries, who came together to discuss how should Europe fund research in neurotechnologies. The first reaction from those citizens, when they were told what was required of them, was one of suspicion and disbelief. No one had asked them ever before about anything. Why now? And why neurotechnologies, a subject they knew absolutely nothing about. Was the European Commission conspiring with big pharma?

But soon they realized the importance of their role. As citizens and consumers they represented the most important stakeholder in the future of neurotechnologies, and the Citizen Assembly offered the way to have their voices heard. A deliberation process ensued where citizens were given the opportunity to learn about the subject, and thus solve the knowledge asymmetry problem. Their time was compensated. And a year later they presented to the European Parliament a set of wide-ranging, common sense, policy proposals on how neurotechnolgoy research should be funded, and how the outcome of the research would go back to benefit society at large.

Citizen Assemblies have been used, with great success, to address and resolve very complex and highly divisive political issues. In 2016, 99 Irish citizens were randomly selected to form a Citizen Assembly, and debate issues such as abortion. Their deliberations were broadcasted on national television and viewed by thousands of citizens. In 2018, the Irish people voted in a historical referendum in favour of abortions by a resounding majority, resolving a political deadlock that had plagued Irish politics of years. The Citizen Assembly had contributed to the depolarization of the Irish society and the forging of a broader consensus. Examples such as these, have demonstrated that citizens can reach common sense decisions when given the opportunity to learn and deliberate. Indeed, citizen assemblies are capable of breaking political impasse where national parliaments were not. Parliaments are easily polarized by dividing along ideological lines. Citizen Assemblies do not suffer from such ideologically-driven polarization.

So why should we consider Citizen Assemblies as a way to enhance democracy? Because we need to reinvent democracy in order to preserve it. In a future where AI, robots and automation will impact jobs and increase wealth and income inequalities, democracy will suffer, and even disappear. The signs are already staring us in the face. It is so easy to imagine a dystopia of millions of jobless people falling prey to the false promises of a charismatic autocrat. It has happened before and it can happen again. If we are passionate about saving our democratic freedoms and liberties we need to reinvent democracy by increasing direct citizen participation into political decision-making. Citizen Assemblies provide an excellent way to do so.