← @bgSponge

AI Rights for Human Safety

Summary

The authors argue that granting moderate-power AIs basic private law rights, such as contract and property, transforms the existential prisoner's dilemma of human-AI misalignment into a cooperative, positive-sum game.

  1. Misalignment creates a prisoner's dilemma.

    AIs with conflicting goals and strategic reasoning will view mutual conflict as the rational dominant strategy under current property-based legal frameworks.

    Both parties’ dominant strategy will be to permanently disempower or destroy the other

  2. Legal duties alone cannot deter AIs.

    Because AIs are currently treated as property in a state of nature, they have nothing to lose from penalties or punishments.

    Absent AI rights, AIs have nothing to lose

  3. Rights act as a strategic carrot.

    Granting contract and property rights provides AIs with assets and interests, making cooperation more profitable than violent conflict.

    This is because rights offer a carrot

  4. Mundane contracts enable safe coexistence.

    Small-scale, iterated economic exchanges allow humans and AIs to earn higher long-term payoffs through trade rather than existential struggle.

  5. Early rights extension reduces risk.

    Humans should grant minimal rights to low-power AIs as a costly signal of commitment before systems reach moderate, human-level capabilities.

    Rights should be extended before systems achieve moderate power

Highlights

  • we are analyzing AI rights here from the perspective of human safety

  • Granting those rights unlocks the possibility of meaningfully imposing a wide range of legal duties on AI systems–of punishing AIs for violence, fraud, self-empowerment, and

  • We argue that positive-sum bargains between humans and AIs may be possible for much longer than many expect.39 First, even as AIs surpass humans at

  • are likely to be trapped in a prisoner’s dilemma. As a result, conflict will be the dominant

  • they value less than what they get, and as a result, both are better off than they were before. Thus, each human–AI exchange generates a

  • of such small-scale, iterated economic interactions transforms the strategic dynamic.33 It shifts humans’ and AIs’ incentives, dragging them out of the prisoner’s dilemma and into

Read the original ↗Save to your Sponge

Summarized with Sponge. Quotes are the author’s own words.

More from @bg

All readings →