An emerging area of work
AI & Society
People affected by consequential uses of AI should have a meaningful say in how those systems are introduced, evaluated and governed.
The public question
AI enters decisions.
People need a voice.
AI systems increasingly influence work, learning, services and institutional decisions. Their effects depend on where they are used, what evidence supports them and who can question the result.
Our work asks how communities and institutions can make those decisions together. Human judgment matters both when a system is used and when its rules are set.
A question at every stage
Who decides?
People affected.
From the beginning.
Throughout use.
A proposed approach: affected people help shape the purpose, safeguards, evaluation and human oversight of AI, with a continuing route to challenge its use.
A use is proposed.
Is AI appropriate here?
Begin with the purpose, the people it serves and the alternatives. Some decisions may be better made without AI.
With the people affected
Affected people are identified.
Whose experience counts?
Include people who use the system, people subject to its decisions and people who may be excluded. Participation begins before deployment.
With the people affected
Use and safeguards are defined.
What rules should apply?
Affected people help define acceptable uses, boundaries and what must be disclosed. Safeguards should reflect the setting and its risks.
With the people affected
The system is evaluated.
What evidence is enough?
Examine performance, limits and effects on different people. Independent evaluation should test the claims that matter in the intended setting.
With the people affected
Human oversight is established.
Who can intervene?
Make responsibility visible. Identify when human review is required and who has the authority, information and time to act.
With the people affected
Use remains open to challenge.
How can people question or stop it?
People need a clear route to question decisions and seek review. Evidence from use should inform changes, pauses or withdrawal.
With the people affected
Our approach
Develop the evidence.
Make room for judgment.
We are examining participatory governance, community-defined safeguards, transparency, independent evaluation and meaningful human oversight.
This is an emerging area of work. We are developing questions for research, convening and applied learning, including how people can challenge a deployment and how institutions should respond.
Connected work
One Foundation. Related questions.
Health & Place
Health access, place-based evidence and public-interest infrastructure remain central to the Foundation. This work examines access, equity and the systems people rely on.
Explore our workAI & Society
How should people and institutions make consequential decisions about AI? The SozoRock AI Lab contributes the applied-learning dimension: real work, human judgment, verification, privacy and responsible use.
Research & Public Systems
The SozoRock Global Institute connects research, publication and convening around assurance, governance, access and systems. AI and society extends those research questions.
Explore the Institute