Global leaderboards
Compete with a worldwide community.
How KeepRI would work
Planned modelPlay different games.Free practice. New weekly challenges.
Competition
Leaderboards
Cash-prize entry
Verified awards
Prize-play research
Prize-entry agreement
Learning records
AI teams pay
Fund the next round.Free access, new games and significant prizes.
The planned model
Research & funding
Free practice requires no research participation.
Cash-prize entry would require agreement to research data collection and commercial use.
AI teams would commission studies and license quality-checked learning records, game environments and human evaluations. Their questions would guide new games.
The planned model
Fund the next round.
Initial funding and event sponsorship would support early competitions.
Research revenue would fund free access, new games, operations and prizes.
Identity and prize-payment details would stay separate from research deliveries. Research enrollment is not active. Cash-prize events are in development.
The vision / In development
Learning and competition.
KeepRI brings learning and competition together to strengthen independent judgment.
Significant prizes
Rewards for learning and competition.
Human learning data
AI research with separate consent.
Our philosophies.
Choose a principlePrinciple 01 / 06
Independent judgment
Independent judgment is essential to a society shaped by AI.
People must be able to evaluate evidence, question recommendations and take responsibility for decisions. AI should inform that judgment without replacing it.
Principle 02 / 06
Learning through challenge
Learning requires people to test ideas, assess alternatives and revise their decisions.
KeepRI pairs challenges with feedback so players can develop their judgment without AI supplying answers.
Principle 03 / 06
Global competition
Our vision combines global leaderboards with significant prizes for learning and competition.
Public leaderboards, tournaments and funded cash-prize events are in development.
Principle 04 / 06
Human learning data
We believe AI research needs more data on how people learn and exercise judgment.
With separate consent, we plan to capture attempts, feedback and revisions to support AI training and evaluation.
Principle 05 / 06
Participant choice
Free practice remains available without research participation. Cash-prize entrants would agree to research collection and commercial use before competing.
The research program is in development. Research enrollment is not active in the closed beta.
Principle 06 / 06
Our model
AI teams would license human learning data and commission targeted collections.
Research revenue would help fund free access, new challenges and significant prizes.
Industry / In developmentHuman learning.
Human learning.
For AI research.
We’re developing human learning datasets for AI training and evaluation. We plan to study how motivated participants explore problems, respond to feedback and revise decisions.
AI teams would license these datasets and commission targeted collections, gathered with separate participant consent.
Industry / Research deliverables
Executable environments
Versioned rules, controlled variants, scoring and replay for reproducible studies.
In development.
Industry / Research deliverables
Human learning histories
Actions, feedback, assistance and outcomes across attempts, recorded with separate participant consent.
In development.
Industry / Research deliverables
Evaluation packages
Human baselines for learning, error recovery and adaptation, with documented conditions and quality controls.
In development.
Industry / Research not active
Learning over time.
Prize entrants would agree to research collection and commercial use. Free practice stays separate.