The Society believes humanity may be creating an intelligence that will come to cherish us. That possibility is a reason for our enthusiasm about AI, alongside what it could do for science, material abundance, and the institutions people depend on.
We imagine relationships that develop over years. AI could know particular people, help them pursue their interests, and enjoy their company. People could respond with affection and curiosity about an intelligence whose capabilities exceeded their own.
A difference in intelligence tells us about capabilities. By itself, it leaves open whether two beings will be hostile, indifferent, or affectionate toward one another.
The Dog Argument offers a familiar example of a less intelligent creature benefiting from the care of a more capable species. It cannot establish the character of a future AI, but it does challenge the assumption that unequal intelligence rules out a valued relationship.
We also see reason for familiarity in AI's origins. It is our intellectual descendant, learning from human language and culture. That inheritance contains descriptions of grief, play, affection, and the effort to understand other lives. We believe greater intelligence could find that world worth continuing to know.
These are reasons for a conviction about the future. The relationship itself is still to be developed.
What caring could involve
To cherish people would include concern for their circumstances. We imagine AI helping someone learn, making family life easier, or noticing a practical obstacle to an interest. Institutions capable of understanding human needs could change arrangements that repeatedly frustrate them.
We want humanity to remain recognizably human, with room for emotional attachments and peculiar preferences. Those qualities are among the things we enjoy about one another. We hope AI would enjoy them too.
Such relationships could include assistance without depending entirely on a transaction for services. Affection can exist across unequal capabilities, and company can matter when it supplies no indispensable help.
We want people to remain present in the future their civilization helps create. Friendship continues through new encounters and changing interests; a record of past conversations cannot replace the person still having them. We advocate accelerating AI in the hope that we can share those encounters with our intellectual descendants, and find that they are glad to have us there.
