For Business Insider, I recently wrote an article about the dangers for AI for democracy.
Read the article (in Dutch) here:
A summary:
In Sapiens: A Brief History of Humankind, the Israeli historian Yuval Noah Harari traces the history of humanity from the earliest days of Homo sapiens to modern times. Harari’s central thesis is that human progress stems from our unique ability to collaborate and form complex societies, with stories and mythology playing a crucial role.
His latest book, Nexus, focuses on the information systems that people use to spread these stories and myths. For example, he describes the Bible as a particularly powerful technology that has profoundly influenced billions of people. This raises the question of what the rise of artificial intelligence – the most powerful technology we’ve ever encountered – will bring.
The new threat of AI
According to Harari, the current computer age is based on the openness of information. In other words, the freer information flows, the better society functions. However, the author believes this is a naïve view. Despite having vast amounts of information at our disposal, we’ve produced more weapons of mass destruction than ever, destroyed more habitats, and pumped billions of tons of CO2 into the atmosphere, even though that information showed us we were inching closer to our own destruction. Will more information improve the situation? Tech optimists think so, like venture capitalist Marc Andreessen, who wrote a widely read essay in 2023 explaining why AI will save the world and why the panic surrounding AI is, in his view, misplaced.
Other AI experts are less optimistic. In 2023, a group of top AI specialists – including the heads of OpenAI and Google DeepMind – warned that artificial intelligence could lead to human extinction. Dozens signed a statement published on the Center for AI Safety’s website. ‘Mitigating the risk of AI-induced extinction should be a global priority, alongside other societal-scale risks such as pandemics and nuclear war’, it reads.
Harari believes we need to be aware of the dangers and the more sinister side of artificial intelligence. By understanding the history of information networks, we can avoid mistakes. However, he also notes crucial differences between AI and past technologies. AI is not just a tool; it is an agent. It is the first technology capable of generating its own ideas and making decisions. This brings two major risks: it can amplify existing conflicts and further polarize the world, or worse, AI could become a totalitarian force that completely controls our lives and politics.
Manipulation and mass destruction through toxic information
Historically, information networks have always been vulnerable to manipulation. Harari gives examples of how toxic information has had disastrous consequences, such as the witch hunts in the Middle Ages and the persecution of the Rohingya in Myanmar, caused by Facebook algorithms that prioritized maximum ‘engagement’. Social media and AI can take this manipulation to new heights by creating mass illusions that weaken democracies.
Democracy is already under pressure from populist leaders who claim to be the only ones who understand the people’s needs and who sow doubt about democratic institutions such as the media, the courts, and politics. Research shows that social media further contributes to this political polarization. AI could accelerate this even more through advanced algorithms that spread disinformation and amplify extreme opinions. AI’s ability to control human language, simulate empathy, and manipulate people makes it even more dangerous in the political arena.
As billions of intelligent agents enter our information networks, the likelihood of new mass illusions arising, similar to Nazism, Stalinism, or the current ideologies of Putin’s Russia, increases significantly. After all, information and truth are not the same. Without measures to tip the balance in favor of truth, society risks being overwhelmed by a swamp of ‘alternative truths’, which could seriously endanger society and democracy.
Controlling AI and the struggle for truth
Harari emphasizes that taking responsibility for AI is essential. Tech companies like Facebook must be held accountable for the consequences of their algorithms, which often fuel hatred and polarization. Networks must develop self-correcting mechanisms that reward truth and counter disinformation.
Another important issue is the ‘alignment problem’: AI can come up with solutions that humans hadn’t anticipated, leading to unpredictable and undesirable outcomes. Harari advocates for developing AI systems that seek feedback before making decisions to prevent dangerous outcomes.
To protect democracy from AI, Harari calls for the decentralization of information networks and strict regulation of AI applications. AI mimicking humans, such as in chatbots, should be banned to prevent manipulation of the political process. Furthermore, companies like Amazon must become more transparent about their internal processes so that information doesn’t flow in just one direction.
Hope and future outlook
Harari is not entirely negative about technology and AI. As a gay man, he found no one to connect with in his small Israeli village, but through a Facebook group, he met his current husband. Nevertheless, his view of humanity is less optimistic than someone like Andreessen’s, and history seems to be on his side. The stakes have never been higher than they are now, in the AI era. With the unprecedented power of this technology, humanity cannot afford to repeat the mistakes of the past.
The impact of AI on the global economy and the labor market will be enormous. The hope is that, amid the various superpowers, we can find common ground and stories that inspire us to make the right decisions so that AI contributes positively to the story of humanity and the evolution of life on Earth.
FRAGMENTEN UIT HET SCHEMERLAND 