In his latest book, Nexus, Yuval Noah Harari highlights the issues of disruptive AI and populist fascism. Specifically, he critiques the naïve view that a free market of information yields the truth because with populism “everyone has their own truth” (p xvii, xxiv). He argues that we need “strong self-correcting mechanisms” (p 404).
Harari elaborates that according to the naïve view, the solution to “misinformation and disinformation is more information”, i.e., more free speech that would expose “falsehoods and fallacies” (p 10-11). However, information overload can crowd out the truth and even technologically advanced economies end up believing in “lies, errors, and fantasies” (p 17). He adds that repeatedly telling a fake memory results in people believing it as genuine (p 24). This is how myths are perpetuated.
Harari disagrees with the Marxist view that wars are fought for the economic interests of the elites instead of “religious, nationalist, or liberal ideals” (p 29). This is because if it’s about power struggle then there is no point in talking to people who disagree with us, and only violence would persuade others to relinquish their privilege (p 30). He explains that Nazis came to power when people came to believe in their harmful narrative (p 31).
Harari states that truth does not resolve political disagreements, and instead “fictional stories” hold “human networks” together (p 33). This is why populist leaders perpetuate national myths of past glory and brush aside inconvenient facts (p 34). For instance, the narrative of Jews as victims sidelined the Catastrophe faced by the Palestinians (p 40-41). However, he clarifies that patriotism is not about hate speeches against minorities but rather about paying taxes so that the collective can have “a sewage system … education, and healthcare” (p 42-43).
Harari argues that democracy allows for self-correcting mechanisms, transparency and accountability, which is undermined when algorithms make decisions
Harari argues that even “free and fair” elections do not guarantee democracy, as it is not the same as “majority dictatorship”, and that populism undermines democracy (p 122, 132). Populist leaders maintain the façade of democracy with elections but undermine countervailing institutions, as they pack the judiciary with loyalists, control universities, and shut down independent media (p 123, 132). The government smears opposing leaders, journalists, and academics as anti-nationals, jails them, or fires them from their jobs (p 123). Populists claim that they alone represent the people and blame “all problems on foreign enemies or internal traitors” (p 131, 180). Harari clarifies that democracy is not majority rule but “freedom and equality for all” (p 124).
Harari notes Facebook algorithms instigated “violence, hatred, and discrimination” against the Rohingyas in Myanmar through fake news on “imagined atrocities” and the false narrative that they would replace the 90% Buddhist majority (p 196). He blames algorithms for the ethnic cleansing of the Rohingyas, as unlike the printing press or the radio, they actively make decisions (p 197, 200). He also notes that surveillance technologies are used with draconian laws and military presence in the Occupied Palestinian Territories, Xinjiang, and Kashmir (p 241). Additionally, he argues that it is pointless to argue against a “propaganda bot”, as it would simply hone its arguments with sustained interaction with us (p 210).
Harari expresses concern that automation can undermine democracy by destabilising the job market (p 316). He warns that computers can replace humans even in jobs that require creativity and emotions, as they are based on pattern recognition that can be subjected to automation (p 318). Moreover, computers have the advantage that unlike humans they are not preoccupied with their own feelings and are free from “stress and financial worries” (p 319, 320). He adds that AI and automation challenge poor countries, as the value of unskilled workers declines (p 374).
Harari argues that democracy allows for “self-correcting mechanisms”, transparency, and accountability, which is undermined when algorithms make decisions, as in the case of job hirings, welfare benefits, and loan approvals (p 331). He adds that democracy ceases to function when people cannot understand and challenge decisions made in a “black box” (p 333). Moreover, when people can no longer make sense of the world and get overwhelmed by information overload, they become susceptible to conspiracy theories and support dictatorship to obtain some certainty (p 334, 343).
Harari mentions the dangers of AI when a dictator confers upon it the power to launch a nuclear strike, when terrorists use it to inflict a pandemic, and when fake news undermines trust (p 361-362). He also expresses concerns on giant corporations turning the globe into “data colonies” and small countries becoming dependent on the AI systems of Big Tech (p 370). He adds that data extracted from the world will flow to the imperial hub and algorithms on autonomous vehicles and disease diagnosis will be exported to data colonies (p 372). Finally, he argues that it is easier to hide an illicit AI lab than an illicit nuclear reactor, so that it is more difficult to agree on AI regulation (p 387).
Overall, Harari argues that free speech does not result in the truth, repetition of propagandistic points can perpetuate myths, populists in power undermine democracy by attacking countervailing institutions, algorithms can inflict oppression, AI can destabilise the job market and undermine democracy, and it is difficult to regulate AI. His key solution is to strengthen self-correcting mechanisms, which means robust checks and balances, transparency and accountability in a democracy.