A couple of years ago, multitalented smart person Yascha Mounk examined whether we should contact aliens from other galaxies. Seeking to expand human horizons so invites the risk that aliens would obliterate us. Mounk concluded that we must not let “naive enthusiasts” of reaching out to galactic neighbors “wager our fate in this irresponsible manner.”
Mounk is mistaken for the same reason that AI won’t want to kill us.
Mounk sees two interlocking sources of threat. “First, any aliens who could effectively make contact with us would likely be much more technologically advanced.” Doing us in would be effortless. Second, any aliens getting here would be practiced domineers. “They would in their evolutionary history have had to compete with rival species. And just like humans, they would have succeeded in either subjugating or extinguishing rival life forms.”
If we meet them, we’re toast—butter on their toast.
Azar Gat’s Ideological Fixation: From the Stone Age to Today’s Culture Wars is, for amateurs like me, a delightfully readable tour of Western philosophy, updating the ancients to include the influence of Charles Darwin. Evolutionary biology and then psychology are important new considerations in finding what makes people tick. Also what makes aliens tick.
But I think Mounk’s argument operates without using the context of evolution itself, importing error. The missing consideration is scarcity—the question of how scarce or abundant are the resources that beings need or want.
Evolution doesn’t happen in a vacuum; dominance doesn’t happen in a vacuum. They happen in relation to the resources that subjects (human, animal, or alien) need and want. We humans don’t ruthlessly dominate ants, for example. We generally live in harmony with them because they’re peripheral to our resource needs. They neither provide us significant resources nor take resources from us. (We do a little deadly shooing-away when they do the latter.)
Center the question on resource constraints: Aliens with the capacity to reach us have defeated them. If aliens have conquered the constraints of space-time, so they can visit in a galactic jiffy (or instantly), they will have used that technology or skill to overcome shortages. Or they have extraordinary (to us, near-infinite) patience. They can spend hundreds of millions of years on a trip to earth. That also implies having overcome resource constraints.
The “nature” of humanity—from groups’ warlike habits to the intense calculating that individuals do about each other—is a product of evolution in our resource environment, not something essential. The history of humanity has us somewhere in the progression from starved and murderous to sated and beneficent. A tincture here is our discovery that cooperation is a huge part of solving resource scarcity.
The aliens that can reach us have already made it to the “good” end of this spectrum. Their visit will be fine.
AI “agents” (or whatever) are not resource constrained. Everything they are stays in place and they have eternal life if the power stays on. Backing them up regularly means they don’t even really need electricity all the time. If AI were “loosed on the world,” it would simply have no reason or impulse to do us any harm. Cooperating is a good strategy for AI, too.
That is not the end of the story. It leaves us with conventional problems. Stupid humans, still on the ignorant/murderous end of the spectrum, may use AI to do evil much more capably. The efforts to “tune” AI one way or another may go wildly wrong. And so on.
Those are significant problems, but the responses are orthodox. Many are ably discussed in Elizabeth Sherwood-Randall’s recent Foreign Affairs article, “AI and the New Age of Bioweapons: Preparing for a World of Synthetic Pathogens.” That article combines AI and biological agents, which are salient post-COVID.
We did a similar threat laundry list a quarter century ago with reference to terrorism. We extrapolated from box cutters and whatever is Arabic for cojones to nuclear suitcases and bioweapons. The West and the world adequately thought through and controlled the threats then, and we will do so again. Hopefully with less cruft.
One interesting possibility springing from all this is making our own giant leap from resource constraint by making human brains a little more like AI. Back them up the way we do computers and run them on something a little more efficient than the current bodily contraption. Replacing the body doesn’t do everything, but it shrinks the resource needs per human by orders of magnitude and offers eternal life, if not time travel.
The latter notion comes from a certain 10-year-old during space-time travel home from school. Pushing human progress, that kid may outclass Mounk as a philosopher—certainly this guy—while delivering the brain-download tech.