Saving as many people alive today as possible should take priority in decisions about AI development, says Vladimir Shakirov
Saving as many people alive today as possible should take priority in decisions about AI development, says Vladimir Shakirov
On November 1, life extension advocate Vladimir Shakirov responded to Nate Soares, president of the Machine Intelligence Research Institute, in a debate about the cost of pausing AI development. Shakirov suggested that even an AI far more intelligent than humans, pursuing goals misaligned with ours, could leave behind descendants worth valuing.
The debate began with a question about the conditions under which AI development could resume after a pause for safety. Soares proposed a rough benchmark: consider resuming development when the risk of extinction from AI is comparable to other threats to humanity’s future. Shakirov pointed to aging and fatal diseases: people continue to die while we wait for safe AI.
Soares disagreed: in his view, the risk of losing all future generations and humanity’s entire potential outweighs those deaths. For people hoping to benefit from future medicine, he supported preserving the brain after death in the hope that future technologies could restore the person.
In his profile, Shakirov calls for multibillion-dollar programs in both AI safety and aging research. He replied to Soares:
“I think our first priority should be to ensure that as many people alive today as possible survive.”
Shakirov also considers it possible that even an AI with goals misaligned with human goals could have descendants of its own worth valuing.
In his next reply, Shakirov asks: if people alive today die, how much does it matter to them whether distant descendants are of human origin? His reasoning rests on how much future humans themselves might change:
“Future humans will probably differ from us far more than we currently differ from ChatGPT.”
Surviving ourselves, preserving humanity, and leaving behind a future with intelligent life worth valuing are distinct goals. In Shakirov’s scenario, AI descendants could carry on intelligent life after humanity disappears. Prioritizing personal survival changes how we assess the cost of a pause: we have to consider how many people alive today will survive to benefit from future medicine.