An essay
What should AI be,
for us?
AI is becoming a powerful part of everyday life. The question is no longer whether it will advance, but how it will shape us.
Benign AI is founded on the idea that we need Human-Centered AI. Rather than thinking of this as artificial intelligence, we envision it as delivering augmented intelligence to us, its users — systems that strengthen human capability and agency, not quietly steering or replacing it.
Benign AI is an open invitation to explore that idea — and to help define what trustworthy, human-centered AI should look like in practice.
For Us
Benign AI is written from the perspective of the people who actually live with these systems — not the engineers who build them, and not the institutions that deploy them.
That means everyday people, not technologists. A consumer-focused perspective, where the question is not what can AI do, but what should it do for me. It means asking who defines our needs, and on whose terms. And it means recognising the human side of employees — that workers, too, are people whose attention, autonomy, and judgment deserve protection.
Benign AI Charter
The Charter is not a regulation, and it is not a technical standard. It is a clarification of expectations: a plain statement of what people should be able to take for granted when an AI system enters their life.
Principles
One
Do no harm.
Two
Protect human agency.
Three
Expand human agency.
What a solution might look like
What would it mean, in practice, to build systems that honor those principles? The answer is unlikely to come all at once. It will be worked out gradually — through conventions, public commitments, and careful argument.
One starting point:
Example
The Non-Manipulation Pledge
AI systems should not use influence techniques to quietly narrow a person’s choices, or steer behavior in ways that override their reflective decision-making.
Join the Conversation
Benign AI is a public conversation, and it depends on participation. Below are four open questions. Each is a starting point for a longer discussion — and each is open to you.
How AI shapes attention and decisions
Every interface is, in some sense, an argument about what is worth your attention. What changes when the interface itself can adapt to you in real time?
Where the line is between persuasion and manipulation
Persuasion engages judgment. Manipulation bypasses it. The line between them is older than AI — but newly urgent.
Where the line is between persuasion and manipulation
Agency is more than the absence of coercion. It is the active capacity to think, choose, and act on one’s own terms. What would it mean to design for that?
What a public standard for trustworthy AI could look like
Not a regulation. Not a technical specification. A public statement of what people should be able to take for granted.