Digital Humanism Summit on Generative AI and Democratic Sustainability

Roger Berkowitz
I spent last week at the Digital Humanism Initiative Summit on Generative AI and Democratic Sustainability in Vienna. Over fifty scholars, computer scientists, regulators, and activists from more than 20 countries met for two full days to think together about the threat AI poses to democracy. Watch the recorded public presentation here, and read our provisional statement below.

Concern about generative artificial intelligence as a future existential threat to humanity should not distract us from the immediate dangers at hand. Consequential elections are coming up in countries around the world, including the United States, South Africa, Taiwan, and the European Union. The integrity of elections and public discourse is already being challenged by social media algorithms that propel and amplify fake news to foster anger in the name of profit.  Generative AI is the next step of the digital, information, and automation revolutions, but it is also a radically new and potentially disruptive technology. The production of fake news will now be joined by AI-produced  counterfeit “human” deep fake speeches and videos. Simulated humans will speak to the concerns of individual voters, explaining why they love one candidate or why they hate another. All of this will make it difficult for the public to determine what is real and what is AI-generated. The combination of social media, recommender algorithms, and generative AI will create a toxic mix that will further undermine the shared public world that is needed in a democracy. 

As generative AI turbo-charges the already polarizing avalanche of misinformation, the likely result will not be the victory of one false narrative over another. The true threat AI poses to democracy is the radical loss of a common world, the  multiplication of virtual and fake worlds, and the distrust of all institutions. The  inability to tell fact from fiction will leave people in a state of radical doubt, isolated and afraid, seeking coherent ideological fictions to give meaning to a chaotic world.


We believe that unregulated generative AI is a clear and present danger to democratic sustainability.  The imminent problem is not super intelligent robots taking over the world, but the threats to human individual and political freedoms posed by the deployment of simultaneously exciting and yet potentially dangerous new technologies. We need to address the full range of AI challenges and, in so doing, the public’s voice must be at the table, not only those of the already powerful. Committed to the  ideals of thoughtfulness, transparency,  and human autonomy, the Digital Humanism Initiative advises:

• Governments and international organizations that convene to talk about the challenges of AI regulation should focus on AI and democratic sustainability alongside the question of the concentration of AI power in large technology companies. 
• While regulation is important, so too is individual liability for those who use social media, algorithms, and AI to subvert democratic discourse. To protect a public world that is verifiable and reliable, we call for the removal or reform of liability shields for large content-producing corporations, such as those provided by Section 230 of the 1996 Communications Decency Act in the United States.
• Aiming to involve diverse and pluralistic voices in the debate about our common future, we pledge to organize participatory processes such as digital deliberative methods and in-person citizens’ assemblies designed to engage and empower publics in finding an emergent consensus on how to respond to the challenges of generative AI to ensure democratic sustainability. 

We invite individuals and organizations to join the Digital Humanism Initiative as impact partners. Contact [email protected]