Humanity still has a sayYour trajectory
BALANCE
Competition inside shared rules.
SHAREPROTECTBALANCE
5 decisions · anonymous · hasasay.com
A question for everyone
What we reward AI for helps shape how it behaves.What should we reward it for?
During training, rewards help shape a model’s responses and decisions—and the future we create with it. Training data and design matter too. The future takes shape through people, AI, discoveries, and choices interacting. Explore five questions about the incentives we build around AI, and add your perspective.
Your answers contribute to this discussion. They do not directly train an AI model.
Having a say is part of what happens next.
We can imagine many futures; our choices help shape the one we live. You do not need to predict it or be an expert to participate. Your answers add your perspective to this public discussion.
5 questionsAdd a commentSee the live view
Question 01 / 05
Humanity still has a sayYour result is an artifact.
Keep it, share it, or ask what it implies.
Want to change an answer?
Your choices stay editable on this device.
One result is useful. A growing public signal is more useful.
Next, add one thing you want people building AI to remember — or jump straight to the live view.
Add your voice
One thing AI builders should never forget.
You can participate even if you skip the five questions. Pick a theme and, if you want, leave your own words. No name or email required.
Your selected theme is counted in the public aggregate. If you write a note, it is stored for review but is not published automatically. Please do not include personal information.
Live public signal
Starting from zero is part of the experiment.
These are real counts from this project. Nothing is seeded, inflated or hidden because the number is small.
This is an open participation project, not a representative scientific survey. The methodology improves as the project grows.
0
Completed experiments
0
Voices added
Where people land.
Five-question result mix
Live anonymous aggregate
What people ask AI to protect.
Themes chosen in Add Your Voice
Live counts
What this data is: voluntary participation from visitors to HASASAY. What it is not: a statistically representative poll of humanity. Poll answers are stored as aggregate counts. Voice themes are counted; optional written comments are stored for review and are not shown publicly by default.
Why collect the signal?
A public picture does not exist until people make it.
The goal is not to claim that a small sample speaks for everyone. It is to build a transparent record that becomes more informative as participation grows.
01 · ADOPTION
What you choose to use.
Tools spread because people adopt them, recommend them and build habits around them.
02 · INCENTIVES
What you reward.
Attention, money and loyalty all teach builders what people value.
03 · EXPECTATIONS
What you refuse to normalize.
Norms change when enough people say a behavior should not become ordinary.
04 · RULES
What you ask institutions to protect.
Public expectations eventually become standards, contracts, governance and law.
After you have your say
Explore the questions with AI.
Ask for an explanation, challenge the site’s premise, or explore a different view. You can also take the conversation in your own direction.
The AI receives what you send in this chat. Your quiz answers and voice comments are not shared automatically. “Ask AI about this result” lets you review your result summary before sending it.
This AI knows the ideas behind HASASAY. You can ask it to explain them, challenge them, or explore a different view. It does not speak for participants. Powered by an OpenAI model.
If you want the deeper version
What is this experiment really asking?
The five questions move from goals, to competition, to dependence, to control, to what “winning” should mean.
Why these five questions?
The questions pull in different directions on purpose. They test how you balance broad access, resilience, trust, distributed power and competition. There is no predetermined correct profile. Results are best understood as a participation signal from this site's visitors, not a representative population survey.
What are the three deeper frameworks?
They are lenses, not predictions and not political labels.
CERN · Shared progressCooperate where the benefits are broadly human while keeping real boundaries around sensitive systems.
PRIMAL · Strategic independenceProtect capability and resilience when dependence on others could become dangerous.
HYBRID · Competition inside cooperationCompete intensely, but inside shared structures that keep rivalry from becoming the whole game.
What is HASASAY actually asking?
What we reward AI for helps shape how it behaves. In training, rewards are signals that favor some outputs or actions over others. They are one influence alongside training data and design. Our five questions explore the wider incentives, institutions and human choices that guide AI development. Having a say is about participating in what becomes real, not predicting it. No individual answer determines the outcome. Your answers contribute to a public discussion; they do not directly train a model.
The conversation that started this
“What game will humanity teach increasingly intelligent systems that they are playing?”
This project began as one conversation about who might win the AI race. The question kept changing until the conversation became an experiment.
Read the origin story →
Engagement without surveillance. Your quiz answers are stored in your browser. When you finish all five, the site sends only your five anonymous selections and resulting trajectory to the aggregate poll counter. Chat messages are sent to the site’s AI endpoint so the conversation can work. No account is required.