Disrupting Democracy? How Generative AI Harms and Helps Democratic Societies
Since becoming widely accessible to the public in late 2022, generative AI (GenAI) has been used for positive democratic outcomes, such as increasing inclusion in policymaking, and for more concerning purposes, like potentially influencing elections. Given the potential benefits and drawbacks, IRI’s Technology and Democracy Practice believes it is critical for its partners to understand GenAI, ensuring that governments and civil society are prepared for current and future impacts.
To understand these current and near-term impacts, IRI launched a Generative AI and Democracy Working Group in November 2023. The Working Group brings together individuals from academia, civil society, industry, and policy to discuss how GenAI may affect democratic processes. Topics have included GenAI’s impacts to information environments, the global digital divide, GenAI’s abuse by autocratic actors, and benefits GenAI has already brought to democracies.
This blog will delve deeper into findings from the Working Group, some of which may be included in a white paper summarizing our work to be released later this year. We’ll review concerns that have been raised regarding GenAI, as well as the advantages of this technology. Lastly, we’ll share a bit about what to expect as the Working Group winds down.
Concerns and Threats
A concern often discussed related to GenAI is how it enables disinformation to be produced faster and more cheaply. A key example of this is an experiment where an AI model created a synthetic media environment using publicly-available tools. The story demonstrated how GenAI can develop, share, and amplify fake narratives on its own, with minimal human involvement. Authoritarian regimes, including China and Russia, may take advantage of these solutions to enhance their influence operations. Meanwhile, online platform moderators may find themselves inundated with GenAI-developed content while lacking the support that could help them stem the tide. Ultimately, the misuse of GenAI by bad actors may decrease public trust in the media, undermining the functioning of democracy in the process.
Additionally, Working Group members highlighted how GenAI magnifies the global digital divide. Simply put, many tools and resources are inaccessible to those who do not speak English. AI-powered products available in the Global South are not as developed or refined compared to those in the Global North. With limited access to quality GenAI tools, or opportunities to gain much-needed skills, individuals in these parts of the world are disadvantaged. Many participants were concerned disparities would widen as GenAI becomes more complex and commonplace.
Hopes and Opportunities
However, Working Group members cautioned against accepting the doomerism that shapes discussions around GenAI and democracy. Recent Working Group sessions have explored the benefits GenAI can bring to governments. Right now, AI-powered tools are leveraged to prepare employer contracts, process legal cases, and transcribe policymaking deliberations. They even tackle more complicated tasks, such as helping domestic violence survivors find support. These discussions shed light on how GenAI is being used for good, ensuring that leaders can better assist their constituents.
Civil society can also take advantage of GenAI. One member of the group is part of the TITAN project which has created a GenAI tool that democratic actors can use to boost citizens’ digital literacy skills. Beyond helping organizations reach their broader goals, GenAI can also assist with day-to-day tasks. Products like Microsoft’s Copilot can help non-profits deal with administrative tasks, allowing them to focus more on time-intensive responsibilities. These use cases only scratch the surface of what GenAI could do for pro-democracy actors.
These are a few highlights from our Working Group series so far. Our final session on May 8th will outline actionable steps those in government and civil society could take to guarantee GenAI works for, rather than against, democracy. Takeaways like these will be incorporated into a white paper to be released in the summer of 2024, and it is intended to be a resource for democratic actors navigating GenAI’s short and long-term impacts. We believe multi-stakeholder collaborations like this project are needed to understand, track, and address GenAI’s impacts on democracies, and help us take advantage of lessons learned from the impact of previous emerging technologies on democratic systems.
Top