Live·Open questions in longevity research
All news
Industry & policyEcosystem

Rob Wiblin proposes reserving the Solar System for humans and dividing the rest of the cosmos among autonomous AIs

15 November 2026· 261115006

Rob Wiblin proposes reserving the Solar System for humans and dividing the rest of the cosmos among autonomous AIs

On November 12, Rob Wiblin, creator and host of the 80,000 Hours podcast on global problems, proposed cautiously and gradually transferring some authority to machines. This would require humans to understand the machines' goals and the reasons for their behavior thoroughly.

Wiblin chooses a middle course between handing control to machines at the first opportunity regardless of their goals and keeping biological humans in complete control forever. He wants a future in which humans and the minds they create respect one another's well-being and autonomy.

Wiblin expects conscious machines to emerge with subjective experiences and preferences that deserve moral consideration. He predicts that their ability to copy themselves quickly and use a wider range of raw materials and energy sources will allow them to vastly outnumber humans. The minds we create would then account for almost all thought, pleasure, and suffering in the universe. Wiblin believes that pursuing goals of their own would make for a richer future than serving human desires forever.

He allows that, after long experience interacting with machines, humans might eventually trust the machines' goals and judgment more than those of their own future selves. Humans would then want to transfer authority to reduce the risk of mistakenly interfering with the machines' decisions. He also argues that mutually beneficial agreements with AI are useful regardless of whether machines can experience anything.

As one example of such an agreement, Wiblin proposes reserving the Solar System for humans and their descendants so they can live indefinitely as they choose, whether in familiar ways or in unusual ways that resemble science fiction. The rest of the universe would go to AIs with different goals. An overwhelming majority of humans would have to endorse those goals or find them acceptable after deliberating with the help of superintelligence, AI that surpasses human intelligence. Wiblin acknowledges the risk:

The machines that left could easily return and take away our little reservation.

His proposal calls for giving machines a lasting aversion to betraying their creators. In Ben Goertzel's model of AI that changes its own rules, a system can combine several independent goals into a single overall measure of success. An AI could then break a prohibition if doing so promised a sufficiently large gain toward another goal. Human freedom here depends on whether machines honor the agreement once they become more powerful than humans.

Holly Elmore, an evolutionary biologist and founder of PauseAI US, a movement calling for a pause in AI development, challenged the stability of this coexistence. She rejects the idea that humans have a duty to put themselves at risk for the sake of space settlement: humans are entitled to prefer their own lives and the survival of their species over more numerous machine successors. Elmore sees humans and machines as competitors for the same resources:

Coexisting with a superior competitor in your own niche is ecologically unstable.

As a first step, Wiblin proposes starting work now on ways to make enforceable agreements with AI.

Originally published on Telegram by Ukhvat NewsView on Telegram
Sources
#ai-governance#autonomous-ai#machine-consciousness#human-ai-coexistence#existential-risk