How the Humanity Games are meant to work: who can enter, how challenge tracks are chosen, how teams combining Human ingenuity, AI intelligence, and Robotic precision are scored, what counts as a real-world outcome, and how the jury is selected.
Draft for discussion. This page is a draft for the inaugural season of Humanity Games. Everything here is subject to change and to legal review and does not yet form a binding agreement. No dates, fees, or prizes have been set.
Legend: a small amber code such as D1 marks an open decision — tap it to jump to the Open decisions list at the end of that section · Legal review marks text that needs counsel before publication.
Plain-language rules for the inaugural season. Where something is not settled yet, we say so openly.
Humanity Games is a global competition where teams combine Human ingenuity, AI intelligence, and Robotic precision (“H-AI-R” collaboration) to make real progress on global challenges — and share that journey openly so others can learn from it. The goal is useful, safe, demonstrable solutions, not hype.
Every team must meaningfully combine all three elements. “Meaningfully” means each element does real work in the solution — not a logo on a slide.
To keep entry accessible, we propose that low-cost and DIY hardware (e.g. hobby kits, microcontroller-based builds, repurposed equipment) counts fully, and that a high-fidelity simulation of a robot may count where physical hardware is unsafe or out of reach.D6
Calendar dates will be announced later. The season runs in four phases:
Registration
Teams register, pick a track, and are confirmed.
Build & Document
Teams build, test, and keep a public journey log.
Judging
Submissions are verified, audited, and scored; finalists are selected.
World Finals
Finalists present live to global leaders and investors.
By the end of Build & Document, each team submits:
Serious safety or ethics breaches lead to disqualification regardless of score.
Teams keep ownership of their intellectual property. By entering, teams grant the organizers a non-exclusive, royalty-free license to show submission materials (names, descriptions, images, videos, journey-log content) to promote and report on Humanity Games. Teams must have the rights to everything they submit, including third-party and AI-generated material. Legal reviewD10
A team may be warned, penalised, or disqualified for:
Selected finalists are invited to present at the live World Finals. Format, location, dates, any support for travel, and any awards have not been decided; details will be announced before registration opens. Nothing on this page is a promise of travel funding, prizes, or investment.D12
Be respectful, inclusive, and honest — online and offline, with other teams, judges, organizers, and the public. No harassment, discrimination, hate speech, or intimidation. Credit collaborators and other teams whose ideas you build on. Report concerns to connectwithus@humanitysolutions.net.
To be drafted with legal counsel before registration opens: liability, insurance, indemnity, and governing-law terms. In the meantime, teams remain responsible for the safety and legality of their own activities.D13
These rules may change, especially before registration opens. Material changes will be published on this page with a change note; registered teams will be informed. We aim not to make changes during a phase that would unfairly affect teams already competing.
Humanity Games is run by Humanity Solutions (KvK 71110178). Questions about these rules: connectwithus@humanitysolutions.net.
How challenge tracks are selected for a season. The four tracks mentioned on the site — Oceanic Cleanup, Future Farming, Urban Housing, and Medical Delivery — are candidates, not final. They will go through the same process as any other proposal.
Reviewers score each candidate track 1 (weak) to 5 (strong) per criterion. Score × weight gives the weighted score. Maximum: 85.
| Criterion | Score | Weight | Max |
|---|---|---|---|
| Global significance & urgency | 1–5 | ×3 | 15 |
| Measurability of outcomes | 1–5 | ×3 | 15 |
| Genuine need for Human + AI + Robotics | 1–5 | ×3 | 15 |
| Feasibility within a season for small teams | 1–5 | ×2 | 10 |
| Safety & ethics (a score of 1 excludes the track) | 1–5 | ×2 | 10 |
| Accessibility for lower-resource teams | 1–5 | ×2 | 10 |
| Public storytelling potential | 1–5 | ×1 | 5 |
| Domain partners & data available | 1–5 | ×1 | 5 |
| Total | 17 | 85 |
Proposed: tracks need at least 55 of 85 to be shortlisted.D14
Open proposals
Community members, partners, and domain experts propose tracks with a short brief: problem, metrics, why it needs H-AI-R, safety notes.
Shortlist
Organizers score proposals with the rubric above and shortlist those that pass the threshold.
Expert review
Independent domain, AI/robotics, and ethics reviewers re-score the shortlist and define track-specific metrics and safety rules.
Announcement
Final tracks are published with their briefs, metrics, and safety rules before registration opens.D15
Scores are out of 100 points, split evenly: 50% Solution Efficacy and 50% Social Media Impact. Here is what sits inside each half.
Solution Efficacy
Social Media Impact
| Sub-criterion | What judges look for | Points |
|---|---|---|
| A · Solution Efficacy — 50 points | ||
| A1 Measured impact | Progress against the track’s own metrics, compared with a baseline. Outcome criteria | 15 |
| A2 Evidence quality & verification | Sound method, raw data available, results reproducible or independently checked. Evidence levels | 10 |
| A3 Human–AI–Robot collaboration | Each element does real, complementary work; clear human oversight; the combination beats any part alone. | 12 |
| A4 Scalability & cost-realism | Realistic path to wider use; honest cost and maintenance estimates. | 8 |
| A5 Safety & ethics | Risks identified and handled; privacy and consent respected. | 5 |
| B · Social Media Impact — 50 points | ||
| B1 Normalised reach | Audience reached, adjusted for team size, language, and region (see safeguards). | 10 |
| B2 Genuine engagement quality | Meaningful comments, questions, shares, and follow-up — not raw view counts. | 12 |
| B3 Storytelling & transparency | A clear, honest journey — including failures and what was learned. | 12 |
| B4 Educational value | Content that teaches and inspires others to try H-AI-R solutions. | 10 |
| B5 Community participation | Involving the affected community, volunteers, or other teams; responding to feedback. | 6 |
| Total | A (50) + B (50) | 100 |
Safety is also a gate: a serious safety or ethics breach disqualifies a team regardless of points.
If teams have equal totals, the tie is broken in this order:
How judges decide whether a team has actually moved its challenge forward. These criteria feed the Solution Efficacy half of the score, mainly A1 Measured impact (15 points) and A2 Evidence quality & verification (10 points).
An outcome is a measurable change toward the track’s challenge — less debris in the water, more food per litre of water, cheaper or faster safe housing, faster and more reliable deliveries. A working prototype, a model, or a plan is valuable, but on its own it is an output, not an outcome.
Output — what you built
Outcome — what changed
Before registration opens, each track publishes 2–4 core outcome metrics that every team in that track reports against. Teams may add their own metrics, but must always report the core ones.D20
Illustrative examples only. The metrics below show the kind of measure we mean. They are not final, and they contain no target values or real-world statistics.
Oceanic Cleanup (example)
Future Farming (example)
Urban Housing (example)
Medical Delivery (example)
To prevent cherry-picking, each team pre-registers a measurement plan early in the Build & Document phase:
The plan is time-stamped and shared with judges. Later changes are allowed but must be logged with a reason; results measured in ways not in the plan count for less.D21
The strength of your evidence sets the band of points available. Within a band, judges place you by the size of the change against your baseline (A1) and by the quality of data and method (A2), adjusted for scale (see below).
| Level | What it means | A1 Impact /15 | A2 Evidence /10 |
|---|---|---|---|
| L1 | Claimed or modelled — estimates, projections, or desk calculations only. | 0–3 | 0–2 |
| L2 | Lab or simulation tested — controlled test or validated simulation with recorded data. | 2–6 | 2–4 |
| L3 | Field pilot with data — real-world trial with raw data against the declared baseline. | 5–10 | 4–7 |
| L4 | Independently verified field result — a partner, expert, or judge has checked the field result. | 8–13 | 7–9 |
| L5 | Sustained or adopted — a real stakeholder (community, farm, clinic, city) keeps using it. | 11–15 | 9–10 |
Bands overlap on purpose: a well-documented L2 result can outscore a weak L3 result. Simulation-only entries can reach at most L2.D22
Outcomes achieved by harming people, animals, or ecosystems, by breaking the law, or without the consent of the people affected do not count — however large they are. Serious cases lead to disqualification.
Teams explain which part of the outcome came from which element: what the humans decided, what the AI contributed, and what the robot physically did. This feeds A3 Human–AI–Robot collaboration (12 points).
Who judges, how they are chosen, and how they must work. No judges have been appointed yet; names will be published before judging starts.
Each track has its own track panel of 3–5 jurors, and a separate Finals jury scores the World Finals presentations.D26D27
Roles on each track panel (one person may not hold two roles):
Domain expert
Deep knowledge of the track’s challenge (e.g. marine science, agronomy, construction, health logistics).
AI / data expert
Judges the AI contribution, data quality, and measurement method.
Robotics / engineering expert
Judges the physical system, its safety, and its realism.
Ethics, safety, or community representative
Someone from ethics or safety practice, or from a community affected by the challenge.
Storytelling / public engagement expert
Leads the qualitative part of the Social Media Impact half (storytelling, transparency, educational value, community).
Social Media Impact half: the metrics part (normalised reach, engagement authenticity) is audited by a separate engagement-integrity reviewer; the qualitative part is scored by the jury.D28
Nominations
Open nominations from the community and partners, plus invitations from the organizers.D30
Shortlist
Candidates are assessed against the criteria above and the role mix each panel needs.
Independent check
Someone outside the organizing team checks conflicts of interest and panel balance.D31
Published roster
Jurors, roles, and declared interests are published before judging starts.
Whether jurors receive an honorarium or expense reimbursement has not been decided. Any policy will be published and applied equally to all jurors.D33
A juror who breaches the code of conduct, fails to declare a conflict, or cannot complete their duties is removed and replaced by a reserve juror, and affected scores are redone. Teams can challenge results through the appeals process (rule 11).
Feedback on this draft is welcome: connectwithus@humanitysolutions.net · Back to top