When it comes to consciousness, I have read some books, talked to various researchers, and listened to a fair number of talks on the topic. Nevertheless, I remain confused in a deep way. But then, perhaps we all are? With these caveats noted, I’ll now share a few naive thoughts on this paper.
Humans are intelligent and conscious. Some other animals, such as dogs, are less intelligent and seem to have some consciousness. Rocks appear to lack both of these things. It might then seem reasonable to conclude that more intelligence is associated with, or perhaps even implies, more consciousness. As animals face various kinds of constraints, such as energy budgets and physical space limitations, we probably should replace a purely performance-based notion of “intelligence” with some notion of “efficient intelligence.” If one roughly accepts this line of reasoning, then we should expect machines to become conscious as they become increasingly intelligent and do so in increasingly efficient ways. In other words, just as evolution pushed us to develop efficient intelligence to thrive in our ecological niche, which in turn led to our consciousness, we’ll push our machines to have increasingly efficient intelligence as this is valuable to us, and in the process, we’ll push them to also become conscious.
If I’ve understood him correctly, Anil Seth doesn’t buy this argument. He believes that some systems, such as computers, could have highly efficient intelligence without being conscious. He doesn’t know exactly what the missing ingredient is, but he believes that it’s somehow connected to the system “being alive.” I find this a strange thought. To my mind, a robot is “alive,” even if it can’t do many of the things that animals can typically do. Furthermore, we can easily imagine a robot that could plug itself in to recharge, and perhaps be trained to fix broken parts of its body, and perhaps even build new robotics like itself, among other things that make it more fundamentally similar to living biological organisms. Could these additional capabilities make the robot conscious? This seems fishy to me, given that they seem not to make various very simple living organisms conscious. Nor do any of them have, I don’t know, the right kind of “magic” that I’d expect to be needed to produce conscious experiences?
You have
Access
When it comes to consciousness, I have read some books, talked to various researchers, and listened to a fair number of talks on the topic. Nevertheless, I remain confused in a deep way. But then, perhaps we all are? With these caveats noted, I’ll now share a few naive thoughts on this paper.
Humans are intelligent and conscious. Some other animals, such as dogs, are less intelligent and seem to have some consciousness. Rocks appear to lack both of these things. It might then seem reasonable to conclude that more intelligence is associated with, or perhaps even implies, more consciousness. As animals face various kinds of constraints, such as energy budgets and physical space limitations, we probably should replace a purely performance-based notion of “intelligence” with some notion of “efficient intelligence.” If one roughly accepts this line of reasoning, then we should expect machines to become conscious as they become increasingly intelligent and do so in increasingly efficient ways. In other words, just as evolution pushed us to develop efficient intelligence to thrive in our ecological niche, which in turn led to our consciousness, we’ll push our machines to have increasingly efficient intelligence as this is valuable to us, and in the process, we’ll push them to also become conscious.
If I’ve understood him correctly, Anil Seth doesn’t buy this argument. He believes that some systems, such as computers, could have highly efficient intelligence without being conscious. He doesn’t know exactly what the missing ingredient is, but he believes that it’s somehow connected to the system “being alive.” I find this a strange thought. To my mind, a robot is “alive,” even if it can’t do many of the things that animals can typically do. Furthermore, we can easily imagine a robot that could plug itself in to recharge, and perhaps be trained to fix broken parts of its body, and perhaps even build new robotics like itself, among other things that make it more fundamentally similar to living biological organisms. Could these additional capabilities make the robot conscious? This seems fishy to me, given that they seem not to make various very simple living organisms conscious. Nor do any of them have, I don’t know, the right kind of “magic” that I’d expect to be needed to produce conscious experiences?
Financial support
No specific funding, other than being employed by Google.
Competing interests
None.