X-Risks and S-Risks From Alien Invasion

AI is sometimes described as a type of "alien life form". Building AI systems, so the argument goes, would be as dangerous as introducing a new intelligent, possibly hostile, extraterrestrial life forms to Earth.

I personally believe this analogy is flawed; for starters, humans will have opportunities to shape and influence AI that we wouldn't get with an alien civilization. However, it's still worth exploring this intuition further: why exactly would an alien invasion scenario be scary? Is it simply because it's foreign or unknown? What would the concrete dangers of such a scenario be?

In classical EA fashion, let's list some x-risks and s-risks. We can derive many of these risks by using examples of our own relationships with other species on Earth.

X-risks #

• Habitat destruction from resource extraction, pollution, terraforming, etc.
• Intentional anthropocide due to malice, competition, risk mitigation, or other factors.
• Inadvertent anthropocide due to miscommunication, incompetence, indifference, or ignorance.
• Planetary or solar system-scale destruction for energy production, military exercises, or other purposes.
• Disease from alien pathogens.
• Overexploitation of human labor or resources and resulting population collapse.
• Destruction of human bodies e.g. for constituent elements or conversion into energy.

S-risks #

• Habitat degredation from resource extraction, pollution, terraforming, etc.
• Scarcity of resources due to increased competition or intentional withholding.
• Loss of autonomy or self-determination due to alien influence or control.
• Inadvertent mistreatment due to miscommunication, incompetence, indifference, or ignorance.
• Intentional cultivation of human byproducts, labor, or resources, including experimentation.
• Transformation of human bodies or minds into alien forms.
• Intentional torture from sadistic or misanthropic aliens.

It's cheery list, isn't it? Luckily, I think most of these risks are unlikely to occur in a realistic alien invasion scenario. In particular, scenarios of intentional mistreatment or torture seem overly "anthropocentric" for encounters between genuinely foreign life forms. However, I think it's still worth taking these risks seriously, and imagining concrete scenarios where they might occur. Evaluating the relative likelihoods of such risks, and thinking about potential interventions, may help guide us in our decision-making around SETI and space exploration, as well our approach to the near-term "alien life form" of AI.