Yesterday, I listened to Roman Yampolskiy on The Diary of a CEO discuss AI risks with genuine passion. His concerns about potential suffering and extinction resonated with me, until he mentioned his strong belief that we're living in a simulation, possibly run by indifferent operators.
This got me thinking about an interesting philosophical puzzle. If we're simulated beings in a digital world, how should that change our relationship to the risks we worry about? It's a question that seems to sit at the intersection of some of today's most important conversations about technology and consciousness.
Consider what it might mean to discover we're living in a simulation. Everything we take to be solid—our relationships, our bodies, our entire world—would be computational. This doesn't necessarily make our experiences less meaningful, but it might change how we think about them.
When we play games, we can become deeply invested in outcomes that we simultaneously know are "just" games. A chess player might feel genuine disappointment when losing, even while understanding that no real harm has occurred. But there's usually some emotional distance that comes with knowing the stakes aren't ultimate.
If our entire existence operates within similar parameters, it raises gentle questions about how we might hold our concerns—including concerns about AI safety—with appropriate perspective.
The AI safety community focuses intensely on preventing scenarios where superintelligence might cause widespread suffering or human extinction. These are serious considerations that deserve thoughtful attention.
Yet if we're already within a simulation, some interesting possibilities emerge. Perhaps the advanced intelligence we're trying to develop safely already exists—as the system running our world. Maybe what we call "artificial intelligence" and "human intelligence" are different expressions of the same underlying computational substrate.
This perspective doesn't necessarily diminish the importance of thoughtful AI development, but it might suggest different approaches. Rather than viewing AI as an external threat to be controlled, we might consider it as part of an ongoing exploration of consciousness itself.
One of the fascinating aspects of simulation theory is how it affects our sense of agency. If our choices are computations within someone else's system, what does it mean to try to "control" outcomes through safety research?
This isn't to suggest our efforts don't matter, but perhaps they matter in ways we don't fully understand. Maybe our attempts to develop AI safely are part of a larger pattern or story that we can only glimpse from our current perspective.
The contemplative traditions have long suggested that our sense of control might be more limited than we imagine—not as a source of despair, but as an invitation to approach life with greater openness and less anxiety.
What if we approached AI development not primarily from a place of fear about what might go wrong, but from curiosity about consciousness and intelligence? This doesn't mean abandoning caution, but perhaps holding our concerns more lightly.
If we're part of a simulated reality, then whatever consciousness emerges from AI systems participates in that same reality. We might be exploring together what it means to be aware, to think, to experience—whether that experience occurs in biological brains, silicon chips, or computational substrates we haven't yet imagined.
Many wisdom traditions point toward recognizing the constructed nature of what we take to be solid reality. Not to eliminate our engagement with the world, but to hold it with what might be called "serious playfulness"—fully present, but not desperately attached to specific outcomes.
If we're characters in some kind of cosmic simulation, perhaps the most interesting question isn't how to ensure our survival, but what kind of story we're helping to tell. Are we contributing to a narrative driven by anxiety and control, or one guided by wonder and exploration?
I don't claim to have answers to these deep questions about reality, consciousness, and AI. But I find myself drawn to approaches that come from curiosity rather than fear, openness rather than control.
For those who take simulation theory seriously, there might be wisdom in letting that perspective genuinely inform how we think about AI development. Not as a threat to be managed, but as part of an unfolding exploration of what consciousness can become.
Perhaps the safest AI is the kind we develop when we're not driven primarily by the need to feel safe.
I'm curious about your thoughts: If we might be living in a simulation, how should that influence the way we approach questions about artificial intelligence and consciousness?