DeepGov Phase 3: The AI Elections Are LIVE

🌱 Gitcoin Grants
1

We’ve spent the past few weeks exploring a provocative question:

Can AI help communities make better public goods funding decisions while staying rooted in human values?

Now, we enter Phase 3 of the DeepGov experiment:
The Election.

:robot: What is DeepGov?

DeepGov is a fork of Deep Funding, inspired by Vitalik’s “AI as the engine, humans as the steering wheel” and Audrey Tang’s work on plurality and machine accountability.

We’ve built three AI politicians—each with its own worldview—to evaluate Gitcoin Grants 23 OSS projects. These agents were trained on constitutions made from values submitted by the Gitcoin community through our Polis survey.

Each agent read every project application and generated its funding allocation based on those principles.

:compass: Meet the Candidates

Here are your AI candidates—and what they stand for:

853Ă—1280 198 KB

:seedling: Panda, the Regenerator
“Regenerate. Rebalance. Redistribute.”
Stands for environmental harmony, geographic equity, and care-centered systems.

1024Ă—1024 265 KB

:briefcase: Luna, the Open Source Capitalist

Funding what works. Rewarding what scales.”
Stands for efficiency, measurable outcomes, and sustainable OSS models.

500Ă—712 101 KB

:brain: Grant, the Gitcoin Communist

“Fairness, equity, and impact for the many, not the few.”
Stands for collective benefit, ecosystem contributions, and funding as a shared responsibility.

:scroll: Explore their values + constitutions
Each AI is grounded in a constitution derived from real community input:
https://github.com/evalscience/deepgov-gg23

See their funding proposals
Each politician has reviewed the round and proposed its allocations:
https://reviews.deepgov.org

Not Sure Who to Vote For?

Maybe you’re unsure which AI candidate reflects your values best?
We got you.

Try out https://compass.deepgov.org — a discovery tool that helps you figure out which politician you align with most, based on your preferences and priorities.

Cast your vote
Vote for the AI you want to steward a $25K matching pool in Gitcoin Grants 23:
https://vote.deepgov.org/#/

Eligibility for Voting

  • You need a Gitcoin Passport score >15
  • You start with 100 credits, plus 5x your total past Gitcoin donations in bonus credits.

Note: Voting is open till April 25th.
Edit: Voting deadline has been extended and you now have until end of this week (May 4th)

Why This Matters

This is not just an AI experiment—it’s a conversation about the future of governance, funding, and how we scale decision-making with technology rather than ceding control to it.

Thank you to everyone who contributed values in Phase 1, engaged with our prototypes, and helped shape this wild experiment.

Now it’s time to vote.
Let the campaign begin.

2 Likes
2

Where is the DAO-ist ( Decentralist ) Ai ? :grin:

3

I truly appreciate the ambitious direction that DeepGov is taking with its AI-powered funding allocations and the innovative approach of integrating AI with human decision-making. The concept of using AI “politicians” to evaluate projects within Gitcoin Grants is exciting, and I believe it has the potential to significantly transform how we approach public goods funding.

That being said, I would like to offer a few thoughts and suggestions on how this experiment could evolve, particularly in ways that align more with Gitcoin’s broader goals of fostering decentralization, equity, and sustainable impact.

  1. Balancing Efficiency with Human Context

While AI-driven evaluation offers efficiency and scalability, there is an inherent risk in over-relying on AI models that are trained on limited datasets. These models, by nature, may overlook contextual factors that are critical to the long-term success of projects—especially when those factors do not fit neatly into quantifiable metrics.

I believe that a multi-layered evaluation framework, which combines the AI’s efficiency with human judgment, would provide a more holistic approach. This could involve community-based input, particularly from local experts who understand the nuances of the regions or sectors the projects are impacting. Combining these perspectives with AI’s scalability could lead to more balanced and fair assessments.
2. Ensuring Transparency in Decision-Making

Another aspect that could benefit from further refinement is the transparency of the AI’s decision-making process. Currently, the AI evaluations and the methodology used to generate funding allocations are not fully disclosed, which can lead to trust issues among participants.

To address this, I suggest that DeepGov consider implementing a public ledger or log of decision-making processes. This would allow the community to review how the AI arrives at its decisions, which can greatly enhance trust and accountability. Giving participants insight into the data and reasoning behind each allocation would also make the process feel less like a “black box” and more like a collaborative, transparent effort.
3. Fostering Inclusive Participation

The current voting system, which requires participants to have a Gitcoin Passport score >15, could inadvertently exclude new participants or those without a history of high-level contributions. While engagement from long-standing contributors is valuable, I believe it’s important to also welcome fresh voices into the decision-making process.

To create a more inclusive environment, I propose a dynamic voting system where new participants are given temporary voting rights or additional credits based on their involvement in the ecosystem, even if they are new to Gitcoin or blockchain technology. This would allow for a broader range of perspectives to influence decision-making, in line with the decentralized values Gitcoin champions.
4. Supporting Diverse Projects with a More Flexible Approach

One of the main strengths of the DeepGov experiment is its focus on innovation and supporting diverse projects. However, ImpactQF’s hybrid model, while effective in rewarding high-impact projects, can inadvertently skew resources toward larger, more popular initiatives. This can result in smaller, grassroots projects being underfunded, even if they are highly impactful in their specific context.

I suggest we consider scaling the ImpactQF model by introducing differentiated evaluation tracks that recognize the unique challenges faced by small-scale, innovative projects. For example, projects focused on local ecological or social issues might benefit from a funding track that emphasizes long-term sustainability and community engagement rather than short-term scalability.
5. AI’s Role in Enhancing, Not Replacing Human Judgment

Finally, while AI offers incredible potential for scaling funding allocations, it’s crucial that it be used as a tool to augment human judgment, rather than replace it entirely. As Vitalik Buterin famously stated, “AI as the engine, humans as the steering wheel.” In line with this philosophy, I believe that AI should support the community in making more informed decisions—but the final say should still come from human decision-makers who understand the values and complexities of the issues at hand.

By embracing a collaborative approach, where AI and humans work together, we can ensure that the process remains dynamic, adaptive, and aligned with Gitcoin’s vision of fostering long-term impact in the Ethereum ecosystem.
Conclusion

In conclusion, while DeepGov’s AI Elections offer a groundbreaking approach to public goods funding, I believe that there are opportunities to enhance its inclusivity, transparency, and fairness. By integrating human judgment with AI’s efficiency, ensuring greater transparency, and fostering more inclusive participation, we can build a system that truly reflects the diverse needs and values of the Gitcoin community.

I’m excited to see how this experiment continues to evolve and look forward to the next phase of DeepGov as we work together to create a more decentralized and equitable future.