AI that makes us
more human.

We pledge that our AI will make people more capable, not less necessary, and safer, not more exposed. We will design, deploy, and govern our systems so they expand human skill, opportunity, and creativity; we will keep pace with those who would abuse them; we will put AI capacity behind charity and causes so nobody is left behind; and we will answer for that commitment in public.

Five commitments

1. 01

   Empower, do not displace

   We build and ship AI that adds to human judgment and craft. Where automation removes work, we invest in the people affected with clear paths to new skills and roles, not silence.

2. 02

   Defend people from bad actors

   We treat security as a first-class product duty. We design against misuse, share threat knowledge where it protects the public, and support tools that keep pace with attackers so people and institutions can feel safer, not more exposed.

3. 03

   Invest where humans grow

   We put real resources into education, tools, and causes that expand creativity, opportunity, and human development. Progress we fund should be progress people can feel.

4. 04

   Allocate capacity so nobody is left behind

   We dedicate a meaningful share of our AI capacity (compute, talent, and tools) to charity and philanthropic efforts. Communities least served by today's AI should still gain from it.

5. 05

   Lead in public

   We publish what this Pledge means for our products, our workforce, our research, our security posture, and the capacity we put behind public-interest work. Once a year we report progress in plain language: what we shipped, what we learned, and what still needs work.

Stewarded by Scott Borchetta and Edward Ginis.