Reclaiming Technology for Democracy by Carl Miller, Fellow at the Centre for the Analysis of Social Media

Adapted from Carl Miller’s recently published essay in the new Forum report Leveraging AI for Democracy: Civic Innovation on the New Digital Playing Field. Read the full collection or watch the launch event.  

As the Internet first came into being, some of its earliest inventors saw it as synonymous with democracy. Yet in the past decade, pessimism has set in along with a succession of commentary making the case that networked computers have torn us further apart. In response, a series of urgent efforts are underway to build new information spaces that buttress democracy, including work on new deliberative technologies: systems designed to let people discuss, consider, and ultimately decide at scale and over distance.

Recently, advances in AI technologies, and language processing in particular, are giving a boost to these explorations. Some of the warmest enthusiasm has been for new deliberative processes using latest generation large language models (LLMs). These processes leverage new capacities to automatically summarize text, translate or moderate conversations, or identify opinion groups in order to facilitate tech-assisted deliberations at scale. But questions remain around where these exercises might fit into civic life and democratic practice.

As AI advances make digital deliberation more dynamic and accessible, the democratic community must bundle technological and political innovation together. Although many discussions focus on national-level government, other contexts may actually hold out the most promise for digital democratic innovation.

Globally, AI-assisted deliberative processes could help international institutions connect more directly with publics, potentially filling gaps in global governance by providing novel opportunities to discuss issues across cultural, national, and linguistic boundaries. Within national democratic systems, local governments have tended to be the most experimental in trialing AI-facilitated deliberation. In these settings, LLMs could help organizers to identify small groups with shared concerns who might participate in targeted (perhaps offline) discussions, whether by sorting through viewpoints expressed in larger deliberations or by drawing on the troves of civic data that local governments hold.

Outside the sphere of formal political institutions, connecting the blockchain-based financial vehicles known as decentralized autonomous organizations (DAOs) with deliberative processes might present an entirely new kind of technological innovation for making decisions—then acting on them. Finally, the most promising applications of AI-enabled deliberation may be in the domain of membership organizations—unions, clubs, associations, trusts, societies, and political parties, as well as more casual, less formally constituted social movements and collectives. Efficient, scaled deliberation represents a new way to represent memberships’ views, values, and priorities, while still retaining bottom-up structures.

There are genuine concerns around these new forms of deliberation and decision making, whether about potential manipulation of digital discussion spaces or the uneasy fit between these approaches and our existing ideas of representative democracy. While the challenges are real, however, the last decade has shown that finding ways of making the technology we use everyday support the democratic systems that we want is not optional. It is essential. Bolstering democratic resilience in the new digital age will inevitably require not just new deliberative technologies, but also changed ways of practicing democracy that can make best use of them.

How Venezuela Became a Model for Digital Authoritarianism

A report from the Atlantic Council’s DFRLab details the Maduro regime’s evolving tactics for digital repression, including an extensive physical and digital surveillance apparatus. The Homeland digital ID—developed in cooperation with the PRC-affiliated company ZTE—enables constant monitoring, a topic covered in a recent Forum podcast featuring report contributor Iria Puyosa. The Homeland system ties together different government databases with sensitive personal information and is required to access crucial social and medical services. The authors also document the spread of surveillance technology supplied by PRC-affiliated companies like Huawei across Venezuela. Read the report here.

Governing AI for Humanity

In September, the UN’s High-level Advisory Body on Artificial Intelligence (HLAB) released its recommendations for the international governance of AI. The report highlights the importance of developing international AI governance structures that provide equal opportunities for all and notes the current lack of representation from the Global South in AI governance fora. To coordinate global AI governance, the HLAB outlines seven recommendations, including the creation of an AI capacity development network and a global AI data framework.  Read the report here.

New Tool Tracks When the CCP Blocks Websites

Rest of World reports on a new platform, GFWeb, that tracks censorship online in China. The tool flags sites that have been blocked by the Great Firewall, provides information on when they were taken offline, and enables researchers to find patterns in the Chinese Communist Party’s internet censorship. For example, evidence gathered through GFWeb showed a trend in blocking AI tools used for generating written, video, or other types of content, indicating government apprehension about widespread use of the technology. Read the article here.

Telegram as a Digital Front in Russia’s War Against Ukraine

The Atlantic Council’s DFRLab documents Telegram’s expansion in Russia since the full-scale invasion of Ukraine and how pro-Kremlin actors are using the platform to spread messages to millions of Russians. On top channels, views grew from nearly 16 billion in 2021 to 95.5 billion views in 2022. As pro-Kremlin channels gain viewers, Russian officials have simultaneously cracked down on critical voices, which is particularly concerning as Telegram becomes a primary news source for many Russians. Read the report here.

A Model for Building Resilience to Foreign Information Manipulation

A new publication by Ben Graham Jones and the Doublethink Lab introduces the Taiwan POWER model, which outlines Taiwan’s approach to countering foreign information manipulation and interference (FIMI). Drawing upon the success of Taiwan’s counter-FIMI approach, Jones seeks to provide a model for global resilience to foreign information manipulation. The report highlights five key attributes of response—Purpose driven, Bottom-up, Whole-of-society, Evolving, and Remit-bound—that should guide strategic stakeholder engagement. Read the report here.

AI in Precision Persuasion

A new report from the NATO Strategic Communications Centre of Excellence explores how advances in AI systems have expanded opportunities to conduct targeted persuasion and manipulation online. For example, the report examines how malign actors used AI-generated content to launch a coordinated campaign to undermine the Ukrainian agricultural industry in Europe. To combat the proliferation of AI-assisted targeting, the report recommends expanding monitoring and detection tools, strengthening safeguards around LLMs, and promoting public awareness of the issue. Read the report here.