The Future of Democracy May depend on Open AI
For-profit AI systems are rapidly reshaping our world, but a critical question looms: at what cost to democratic principles? We’re increasingly reliant on technologies developed behind closed doors, with opaque training data, biased algorithms, and a lack of public oversight. This poses a significant risk, particularly when these systems are deployed in contexts vital to a functioning democracy.
The core issue is control. We lack clarity into how these AI models are built – the data they learn from, the safeguards implemented, and the values embedded within them. Even more concerning, we have no voice in shaping these crucial details. Relying on proprietary AI in democratic processes is, in many cases, simply unacceptable.
The Rise of “Public AI“ – A democratic Alternative
Fortunately, a powerful counter-movement is gaining momentum: the development of “public AI.” This envisions AI models and systems built under democratic control, designed for public benefit, not corporate profit. It’s a global trend, and a significant milestone has just been reached.
Switzerland recently unveiled Apertus,arguably the world’s most advanced and fully realized public AI model.Developed collaboratively by the Swiss government and ETH Zurich, Apertus is a game-changer.
Here’s what sets it apart:
* Fully Open Source: The data, code, and model weights are all publicly available.
* Ethical Training: It was trained without using illegally obtained copyrighted material.
* Fair Labor Practices: Apertus avoids exploiting underpaid workers in developing nations.
* Competitive Performance: Its capabilities are comparable to those of major corporate AI models from just a year ago – more than sufficient for many applications.
* Cost-Effective: Apertus demonstrates that building powerful AI doesn’t require multi-trillion dollar investments.
Apertus isn’t just a technical achievement; it’s a powerful demonstration that an alternative to big tech’s control over AI is possible.
AI: A Tool That Can Empower or Enslave
AI is, at its heart, a power amplifier. It can magnify the intentions of those who create and deploy it. This means it can bolster authoritarian regimes just as easily as it can strengthen democracy.The choice is ours.
We must actively steer this technology toward a more equitable and democratic future. Here are four key ways to do that:
- empower Citizen Oversight: Equip watchdogs and litigators with AI tools to enhance their ability to monitor government and ensure accountability.
- Inform and Engage Voters: Enable political parties and election administrators to use AI for meaningful voter engagement and information dissemination.
- Demand democratic Alternatives: Encourage governments to develop and provide public AI options, challenging the dominance of big tech.
- promote Transparency & Auditability: Advocate for open-source AI models like Apertus, allowing for public scrutiny and enhancement.
AI technology carries inherent risks, and we shouldn’t downplay them. However, its potential benefits are equally significant.
The future of democracy isn’t predetermined. It depends on our collective action to ensure AI serves the public good, not just corporate interests.
This essay was co-authored with Nathan E.Sanders and originally appeared in The Guardian.*
Posted on November 25, 2025 at 7:00 AM • 8 Comments
Keep reading