Pair-Wise Voting for Collective Intelligence – Your Guide to Voting
The first round of voting for the BGI Nexus grant funding is almost here! The pilot round has been so smooth, with amazing community support and energy; we’re excited to enter this final phase and for the wide engagement we anticipate.
Today we’ll share an overview of the voting process, tips for reviewing proposals, background into the expert ratings and review process, and as well as the list of proposals eligible for voting. This information will prepare you, the voting participants, to prepare for the selection process.
Voting Opens: 02 April, 2025, 18:00 UTC
Voting Closes: 07 April, 18:00 UTC
Voting Portal: Click here to vote!
Voter’s Guide → We’ve collected a brief overview of the Pairwise Comparison Method, how to vote, and tips for evaluating proposals in a brief Voter’s Guide PDF which can be shared within your network, helping everyone to vote!
Pair-Wise Voting for Collective Intelligence
In the spirit of the BGI Nexus initiative—founded on inclusiveness, transparency, and collective wisdom—we recognized early on that traditional voting methods weren’t the right fit for selecting grant proposals. Choosing a single “best” project, or top few projects, risks turning a deeply collaborative process into a competition or popularity contest.
We believe that everyone contributing to beneficial AI is part of the same movement. There are no winners or losers here—just a global collective working together to lift up the ideas with the greatest potential for positive global impact.
That’s why we’ve chosen pairwise voting as our decision-making method for this grant round. It’s simple, fair, and grounded in the idea that wisdom can emerge from many people making small decisions, together. With pairwise voting, each participant helps prioritize proposals by comparing just two at a time. Over many comparisons, a clear and community-driven picture of our shared priorities emerges.
How It Works
- Sign in and Verify: Voters enter their email address and confirm with a quick double opt-in.
- Review Proposals: Each voter will see 8 pairs of proposals. For each pair, the voting platform will list project titles, overviews, and links to full proposals hosted on Deep Funding.
- Make a Decision: Based on quality, impact, innovation, ethics, and alignment with our shared mission, you’ll select which project you believe should be prioritized for funding (see the tips on reviewing proposals below or the voter’s guide, attached).
- That’s It: The platform handles the rest—showing proposals to different voters, comparing results, and surfacing priorities through collective input.
Each voter sees a unique subset of the proposals—there’s no need to review and rank all 25. And yet, through everyone’s contributions, a community-wide priority list begins to emerge. Even when voters don’t see the same proposals, the algorithm can infer rankings based on overlapping judgments in a robust and gaming resistant way. It’s a bit like magic—but powered by our collective intelligence.
Why Pairwise Comparison Ranking is a Great Fit
Pairwise voting is low-friction, newcomer-friendly, and helps keep the process grounded in thoughtful comparison instead of overwhelming decision-making. It naturally highlights the most impactful and aligned proposals—not because they got the most clicks, but because they consistently resonated in one-on-one comparisons across the participants.
Pairwise voting has been successfully used in many fields—from participatory budgeting to academic research to community governance—because it offers a simple, bias-resistant way to surface collective priorities through small, thoughtful comparisons. (See additional insights on the benefits of pairwise comparison for decentralized decision-making here and here…and a cool demo that helps give an intuitive grasp of its power.)
The community-prioritized proposals from this round will receive funding. Those with lower priority in this round are still valued contributions, and we encourage all participants to continue refining their ideas, and reapplying in future rounds.
Together, we’re shaping a future where beneficial AI emerges not from competition, but from collaboration— and from the collective intelligence of a global, heart-led community.
How to Thoughtfully Review a Proposal
You don’t need to be an expert to make a meaningful contribution—just bring your care, curiosity, and honest judgment. The goal of this process is not perfection, but participation. Every thoughtful vote strengthens our collective insight.
Here’s how to approach reviewing proposals:
1. Read the Full Proposal
Each voting card shows a brief overview—but to get the full picture, click through to read the complete proposal on the Deep Funding platform. You’ll find:
- The proposer’s name and team background
- Project goals and outcomes for social or environmental impact
- Technical details of the project and building methodology
- Timeline, milestones and use of funds
- Alignment with AI safety and ethical values
- Many projects have also provided a brief 3-minute video introduction to their project, included on the proposal page, or have had their Spotlight Day Presentation highlighted. Some projects also have additional links to videos, websites, demos, repositories, etc.
A note on milestones – just as in the Deep Funding grant program, funds for awarded projects will be disbursed only at the completion of each milestone, as demonstrated by a team report upon milestone completion.
2. How to Prioritize Proposals – Round Criteria
When choosing between two proposals, consider how they compare across these dimensions:
- Impact: Does the project aim to solve a meaningful problem and produce broad or deep impact?
- Innovation: Is this a meaningful step on the path to beneficial AI, bringing something genuinely new, or building on existing tools in a novel way?
- Ethics & Safety: Does the project account for possible risks, harm reduction, data use and privacy, and responsible deployment?
- BGI Alignment: How well does the proposal reflect the values of the BGI Nexus—such as inclusivity, decentralization, and human-AI co-evolution? Also, consider alignment with the round theme “Building a Compassionate and Abundant Future”
Use these as guiding questions to help make your choice – if you have difficulty prioritizing between certain options, just do your best and trust the community will all contribute to sorting amongst nuance.
3. Additional Resources
Along with the written proposal, there are
- ▶️ Spotlight Day Video: The presentations from Spotlight Day were recorded; many of the projects were able to attend and highlight their project to the attendees, and answer community questions. The full playlist will be made available shortly on the Deep Funding YouTube channel.
- 💬 BGI Telegram Group: Join the conversation, ask questions, and hear what others are thinking in real time; many of the proposers are in this group, and may be able to answer proposal questions as well! Join here.
Time spent digging into these wonderful projects will be well rewarded with inspiration. Your time and participation helps shape a more just, imaginative, and beneficial AI future for all.
Behind the Scenes: How Proposals Were Reviewed
Wondering how each project earned its expert ratings?
Community-driven funding brings unique challenges—especially when it comes to ensuring that proposals are high-quality, relevant, and ready for review. Without a careful pre-screening process, voters could be overwhelmed by proposals that vary too widely in eligibility or feasibility. (For more context, check out Deep Funding’s blog post on “The Funding Trilemma.”)
To address this, we recruited a cohort of volunteers from the BGI Nexus and Deep Fund membership to deeply review each proposal. Every reviewer focused on just a few proposals, using a standardized evaluation rubric based on the BGI Nexus Grant Round Rules.
Each proposal was scored on:
- Impact
- Innovation
- Safety and Ethics
- Alignment with BGI Nexus and Grant Round goals
These reviews helped ensure only the most aligned, high-quality proposals advanced to the next phase.
Then, proposals that passed this expert review moved to a technical feasibility check. In this stage, experienced developers and software engineers reviewed each proposal to confirm the team’s technical ability and whether the project was realistic to implement. These technical reviews were pass/fail only, with no numeric scores.
The result?
Every project now eligible for voting has:
- Met eligibility, quality, and safety/ethics standards
- Passed technical review
- Been recommended by reviewers as worthy of funding
Now it’s your turn.
Each proposal on the Deep Funding portal includes:
- Impact and innovation scores from community reviewers
- Written review excerpts (anonymized)
- A short summary paragraph capturing the team’s feedback
Explore these insights as you vote, and help decide which projects will lead the way in this first BGI grant round.
What’s Next?
Once voting wraps, results will be carefully processed and verified. The awarded projects will be announced during the week of April 14.
To be the first to hear the results and stay updated on future grant rounds, events, and community initiatives, be sure to subscribe to the BGI Nexus newsletter or follow us on your favorite channels.
Eligible Projects
Finally, let’s take a look at the 25 proposals which have made it to the voting phase. Please help us in congratulating all of these amazing, thoughtful, potentially world-changing projects, and show your appreciation by participating in the voting event this week.
Voting Opens: April 02, 2025, 18:00 UTC
Voting Closes: 07 April, 18:00 UTC
Voting Portal: (poll unit or Tilda site)
Click on any of the eligible projects below to read more, and prepare yourself to vote next week:
- NigerGuard:AI Climate Resilience-Vulnerable States – equiping Nigeria’s most climate-vulnerable regions with AI-powered risk prediction, smart farming, and disaster response tools. Fusing traditional wisdom with modern tech, it’s a digital shield for resilience—protecting people, crops, and futures against extreme climate shocks.
- GlucoseDAO – GlucoseDAO brings open-source, personalized glucose forecasting to diabetics and health enthusiasts alike. By combining community data with AI, it empowers users to predict blood sugar one hour ahead—enhancing safety, freedom, and real-time wellness planning.
- Project Lovelock – Living Libraries Network (LLN) – building mobile “Living Libraries”—AI-powered hubs blending environmental sensors, satellite data, and regional wisdom. These eco-kiosks empower communities to restore land, adapt to climate change, and steward biodiversity through data-driven, locally-grounded regenerative action.
- Autistic AI Safety Framework & AGI Confidant -building an AI ethics and training framework rooted in autistic perspectives—by and for neurodivergent people. Its prototype AGI “confidant” will support autistic users with privacy-respecting, compassionate interactions that reflect spiky ability profiles and emotional nuance.
- Think Inclusion Conferencing App – an AI-driven meeting platform which bridges the digital divide by translating between speech and sign language in real time. It empowers speech-impaired individuals to fully participate in virtual conversations—redefining inclusivity for the digital age with gesture-based interaction and AI voice customization.
- BIONEXUS GAIA: AI for Regenerative Biodiversity – a decentralized AI and blockchain platform for biodiversity conservation. It integrates species identification, habitat mapping, citizen science, and carbon credit verification to enhance ecological monitoring, ethical AI alignment, and regenerative economic models in support of global sustainability.
- Productivity EEG-AI Agents – Your brain’s secret productivity rhythms—decoded by EEG and optimized by AI. This neurotech platform delivers real-time, personalized focus coaching using reinforcement learning, helping users dodge burnout and boost output with science-backed cognitive flow tracking.
- Scale Free Alignment – What if we aligned not just AIs, but the whole system? This proposal fuses biological self-organization and agent identity theory to craft scalable, provable alignment frameworks—offering a bold reimagining of how AI and society evolve together.
- CyberPravda: sybil-resistant service of veracity – a radical truth engine, this project is a decentralized, math-driven system that assigns veracity scores to information without admins or oracles. Designed to prevent “neural network collapse,” it protects AI—and humanity—from drowning in its own hallucinations.
- Acropolis OS – Autonomous Community AGI – building community AI agents that help groups self-organize, learn, and adapt like living systems. By using ontologies and knowledge graphs, it creates digital collective consciousness—empowering resilient, self-governing communities in the face of social fragmentation.
- London Voice – turning local concerns into actionable insights by blending AI, blockchain, and community input. Through a decentralized app, residents vote, share stories, and co-train a large language model that reflects their needs—shaping policy with collective intelligence.
- Ethical AI Auditing – a practice based approach – developing an ethical AI auditing framework using blockchain and citizen science to revolutionize biodiversity conservation, providing global ecological data, transparent carbon credits, and fostering public engagement in environmental protection.
- Fair AI Content Moderation – developing an AI-powered content moderation system that promotes fairness, reduces biases, and upholds freedom of expression, using NLP, fairness-aware algorithms, and Explainable AI with blockchain integration for accountability.
- Stress Quantify – AI-Powered HRV Analysis – bringing clarity to biometric stress data by standardizing heart rate variability across devices and applying adaptive AI, the project transforms vague graphs into personalized, actionable wellness plans—finally making wearables a true partner in mental resilience.
- Aid for Speech Impairment using AI (ASAI) – ASAI gives voice to the unheard by translating slurred or impaired speech—especially from those with dysarthria—into clear text using a custom-trained AI model. It brings dignity and connection to millions, turning AI into a true assistive ally.
- Sign Language Translator AI (SLTA) – using computer vision and deep learning to translate sign language videos into readable text. Built for the SingularityNET marketplace, it aims to boost accessibility, enabling millions of deaf and hard-of-hearing users to communicate more freely in real time.
- AURORA: Decentralised AI for Carbon Credit Trust – building a decentralized AI system that verifies carbon capture performance with live data, predictive analytics, and blockchain auditing. Its goal: to restore trust in ESG reporting and carbon credits with verifiable, tamper-proof transparency.
- ReGenAI – pioneering “Ecological AI” to build a context-aware system that fuses Indigenous wisdom, regenerative design, and climate data to inform urban resilience. It helps humans and AI make decisions that serve not just efficiency, but the living world.
- Gamifying Benevolent AGI: Elowyn – creating a gamified training environment for benevolent AGI through a strategic card game. It couples a deceptive “Moloch AI” with a guiding “Elowyn Tree AI,” enabling players and AI to co-evolve ethical behavior within win-win, complexity-informed game scenarios.
- Integrated Information Decision-Making System – applies Integrated Information Theory (IIT) to quantify systemic coherence (Φ) in complex environments such as cities, AI systems, and supply chains. It provides real-time insights to inform human and AI decision-making for sustainability, ethical governance, and systemic resilience.
- Farmlingua – a multilingual, AI-powered farm assistant delivering real-time, localized advice—even offline—to smallholder farmers. With voice input, NLP, and sensor data, it empowers rural growers with crop-specific insights on soil, pests, markets, and sustainability, transforming agricultural knowledge into food security.
- SleepAI – SleepAI turns any wearable into a clinical-grade sleep apnea diagnostician. Validated by top-tier research, its AI detects disorders over multiple nights at home—making diagnosis cheaper, smarter, and radically more accessible for the 1 billion undiagnosed.
- Human-AI Collaboration Governance Framework Polis.ai – reinventing governance as a participatory partnership between citizens and AI. From blockchain-secured voting to predictive policy simulations, it empowers communities to co-create legislation and streamline civic action through inclusive, transparent, and ethics-aligned AI decision support.
- Benevolent AI Agent Interactions – a modular AI interaction framework that makes agents more empathetic, emotionally aware, and less robotic. Featuring “Chain-of-Emotion” prompts and symbolic logic tools, it helps developers build AI that connects—not just computes—with humans and other agents.
- Valuing Invisible Flows for Collective Evolution – making the invisible visible: this project uses AI to analyze voice notes, reflections, and emotional flows across communities. It reveals how collective thinking evolves over time, offering a new lens for designing regenerative systems and real-time cultural shifts.
Thank You for Being Part of BGI Nexus
The BGI Nexus is more than a funding round—it’s a growing, evolving community of people who care about building a future where technology uplifts humanity. Whether you contribute through ideas, resources, collaboration, or your vote, you are part of shaping that future.
Every voice matters. Every vote counts.
This week, your decisions will help determine which projects receive support—and in doing so, you’re directly influencing the trajectory of beneficial AI.
Whether you’re new to the community or have been with us from the beginning, your curiosity and care are exactly what we need.
We’re honored to do this work alongside you. Let’s build something extraordinary—together.