Our mission
Make safe AI understandable, relevant and participatory.
We translate research without turning uncertainty into certainty, connect people across borders and create practical ways to contribute to public understanding.
Mission / Why PAIP exists
Intelligence is accelerating. Public understanding must catch up.
PAIP exists to give a new generation the language, evidence and community needed to take part in decisions about increasingly capable artificial intelligence.
Independent · International · Human-centredThe challenge
Control becomes harder as capability grows.
A system can be useful today and behave differently when its capabilities, access or environment change. The central challenge is not simply making AI obey one instruction, but keeping its behaviour reliably compatible with human intent under conditions we cannot fully predict.
Our vision
A world where AI progress does not reduce human freedom or agency.
People should remain able to understand, question, influence and correct the systems shaping education, work, institutions and everyday life.
HUMAN
AGENCY
AGENCY
What “Protected” means
We are not protecting AI from humanity.
Protected refers to the human life, freedom, dignity and influence that must remain secure as intelligence becomes more powerful.- Human life and physical safety
- Freedom to choose and refuse
- Dignity and meaningful participation
- Influence over a shared technological future
What we do not claim
Clear limits are part of credibility.
PAIP is not an AI laboratory.
PAIP has not solved alignment.
PAIP does not represent SSI or any other research lab.
PAIP is not a religious, personality or fan movement.
PAIP does not promise financial success.
Principles
The rules that keep the mission ahead of the market.
Mission before market
Commercial attention must never redefine the reason PAIP exists.
Understanding before fear
We explain risk without panic, spectacle or false certainty.
Source before headline
Traceable evidence matters more than a dramatic claim.
Uncertainty is visible
Unknowns and hypotheses are labelled rather than hidden.
People remain subjects
AI should strengthen human action, not turn people into passive recipients.
Correction is mandatory
A credible project must preserve a visible path to correct errors.