From business and government to everyday life, artificial intelligence is reshaping the world, and politics could be next.
While the idea of AI politicians may be unsettling to some, research tells a different story. A poll conducted by my university in 2021, during the early surge in AI advances, found widespread public support for incorporating AI into politics in many countries and regions.
A majority of Europeans said they want at least some politicians to be replaced by AI. Chinese respondents were even more bullish about AI agents deciding public policy, while Americans, who typically favor innovation, were more cautious.
As a philosopher who studies the moral and political questions raised by AI, I see three main paths to integrating AI into politics, each with its own mix of promises and pitfalls. .
Some of these suggestions are more outlandish than others, but when you weigh them, one thing is certain. AI’s involvement in politics means that we are forced to consider the value of human political participation and the nature of democracy itself.
Read more: Generative AI like ChatGPT could help promote democracy if key hurdles can be overcome
Will chatbots become candidates?
Before ChatGPT burst onto the scene in 2022, efforts to replace politicians with chatbots were already well underway in several countries. Back in 2017, a chatbot named Alisa challenged Vladimir Putin in the Russian presidential election, while a chatbot named Sam ran for office in New Zealand. Denmark and Japan are also experimenting with chatbot-led political initiatives.
Although these efforts are experimental, they reflect long-standing curiosity about the role of AI in governance in diverse cultural contexts.
On one level, the appeal of replacing flesh-and-blood politicians with chatbots is quite obvious. Chatbots do not have many of the problems and limitations typically associated with human politics. They are not easily seduced by desires such as money, power, and fame. They require no rest, engage with everyone virtually at the same time, and offer encyclopedic knowledge and superhuman analytical abilities.
But chatbot politicians also inherit the flaws of today’s AI systems. These chatbots, which utilize large-scale language models, are often black boxes and our insight into their reasoning is limited. They frequently produce inaccurate or fabricated responses known as hallucinations. They face cybersecurity risks, require extensive computational resources, and require constant network access. They are also shaped by biases derived from training data, social inequalities, and programmers’ assumptions.
Read more: ‘Cross-hallucinations’: Why AI has a hard time understanding that six-year-olds can’t become doctors or claim a pension
Additionally, chatbot politicians would not be a good fit for what we expect from elected officials. Our institutions are designed for human politicians with human bodies and moral agency. We expect politicians to not only answer questions, but to supervise their staff, negotiate with colleagues, show genuine consideration for their constituents, and take responsibility for their choices and actions.
Without significant improvements in technology or a fundamental rethinking of politics itself, the prospects for chatbot politicians remain uncertain.
Direct democracy powered by AI
Another approach seeks to eliminate politicians altogether, at least as we know them. Physicist Cesar Hidalgo believes that politicians are a troublesome intermediary that AI can finally eliminate. Rather than electing politicians, Hidalgo wants to allow citizens to program AI agents to suit their political preferences. These agents can automatically negotiate with each other to find common ground, resolve disagreements, and draft legislation.
Hidalgo hopes the proposal will unlock direct democracy and allow citizens to participate more directly in politics, overcoming traditional barriers of time constraints and legislative expertise. This proposal seems particularly attractive given widespread dissatisfaction with traditional representative institutions.
But eliminating expression may be harder than you think. In Hidalgo’s “avatar democracy,” the de facto kingmakers would be experts in designing algorithms. They may simply replace one form of representation with another, since the only way to legitimately recognize their power is probably through voting.
Read more: Opening the black box: How “explainable AI” can help you understand how algorithms work
the ghost of the algorithm
More radical ideas include removing humans from politics altogether. The logic is very simple. If AI technology advances to the point where it can reliably make better decisions than humans, what will be the human input?
An argocracy is a political system run by algorithms. While few would explicitly advocate a complete transfer of political power to machines (and the technology to achieve it is still a long way off), the specter of argocracy raises questions about why humans should be brought into politics. We need to think critically about whether participation matters. How do we maintain values such as autonomy, responsibility, and deliberation in the age of automation?
Future direction
The dramatic potential for integrating AI into politics makes this an important time to clarify our political values. Rather than rushing to replace human politicians with AI, today we can focus on tools that enhance human political judgment and close the gaps in our democracy. Tools like Habermas Machine, an AI debate mediation tool, have successfully helped test groups reach consensus when voting on divisive and polarizing topics. We need more innovations like this.
From my own perspective, the future of AI in politics lies not in completely replacing human decision-makers, but in thoughtful integration that amplifies human capabilities and strengthens democratic institutions. If this is the future we want, we need to be intentional about building it.
