McClelland argues that the problem is more basic: we still do not know what causes or explains consciousness in the first place, which means we do not have a solid foundation for testing whether AI has it.
The problem is even more basic than just recognition --- we do not have a solid foundation for building AI that has it.
In general, it's really hard to build software to mimic something that isn't even defined yet.
I hear this a lot but I've never understood why people think it's a deal breaker. You don't need to start from definitions and in fact sometimes that gets it exactly backwards because the point of research is to understand something well enough that in light of the research, you actually can define it, eg Dark Matter. Or, in a different time, AIDs before we knew what it was.
You can have clusters of related case studies that share the observable effects, and reason and research your way to correlations, and investigate those to discover causation and mechanisms, and infiltrate the "black box" of an unknown thing deeply enough that you account for the whole thing itself.
I think progress on consciousness research in humans is advancing impressively, identifying exactly the kinds of pre and postprocessing done to sensory input and areas of the brain associated with conscious activity and brain to machine interfaces are improving all the time.
Granted the hard problem is still hard and must be respected rather than talked past but the point is we're not stuck. Understanding is gradual and you can model phenomena to the degree that they are understood, closing in from multiple sides.
I hear this a lot but I've never understood why people think it's a deal breaker.
Maybe it's a deal breaker and maybe it's not.
At this point, we're groping in the dark. We don't know enough to stay anything for sure. But we're throwing $billions at it based on the pure hope of somehow getting lucky.
For example, we can't even say for sure that consciousness is something that is isolated in the brain. Neurons exist throughout the human body. Consciousness could very well be a whole body phenomenon --- or not, we really don't know.
Does anyone really know if a human body can be realistically mimicked by software at this point? How much energy and computing power would be required to do so? Is there enough on the planet?
Bottom line: At this point, there are more questions than answers. The one thing we know for sure --- tech billionaires are raking in tons of money from AI.
I suppose I don't disagree with the individual observations, but where you lose me is in treating it all like it amounts to some kind of conceptual system crash that precludes progress as a matter of principle. It's not even that I disagree one way or the other so much as it's a matter of having trailed off from the question of the role of definitions in research.
Maybe it helps to consider dark matter. What we have is effectively a placeholder definition based on its observable effects. We don't know if it's WIMPs, axion-like particles, or even some alternative framework for gravity. But we have enough to state meaningful questions about it and iterate toward understanding from a number of directions using a combination of hypothesis, data and experimentation. Finding out what it truly is would be the culmination of research that settles the question rather than something to be stipulated at the start.
So depending on how you look at it, you already have a working definition of consciousness sufficient to organize research, we already have made real progress of the kind that should be impossible if definitions were really dealbreakers, and having "a definition" in the complete sense is something you would never have up until the point the question was settled once and for all, which happens at the end of research rather than the beginning.
I see Wittgenstein mentioned more often in these parts which is awesome, and I think the best Wittgensteinian attitude to adopt here is to turn the tables on this whole question and refuse to agree that there's such a thing as a question of definition that stands between us and research progress.
Because consciousness is not a scientific nor philosophic phenomenon?
Like all references, its symbolic existence exists within the mind of the observer. It's the mistake of reification[1] to forget this authorship and go looking for it as if it exists out there in reality.
I hate to be the bearer of bad news, but the engineer's distaste for soft sciences, and above all sociology, increases the susceptibility of this fallacy. You may be more familiar it's formulation as "the map is not the territory."
Looking for consciousness in the mathematics of ANNs will yield just as much insight as tearing apart the fibers of dollar bills looking for the essence of financial value, distilling paints to find the soul of art, or searching for gender in genetic material. It's a category mistake[2] wrapped in bias. We'd be much more well off learning how the ontology of consciousness is constructed than continuing to apply the hammer of hard science.
How do we know the Cambridge Philosopher is conscious?
McClelland argues that the problem is more basic: we still do not know what causes or explains consciousness in the first place, which means we do not have a solid foundation for testing whether AI has it.
The problem is even more basic than just recognition --- we do not have a solid foundation for building AI that has it.
In general, it's really hard to build software to mimic something that isn't even defined yet.
I hear this a lot but I've never understood why people think it's a deal breaker. You don't need to start from definitions and in fact sometimes that gets it exactly backwards because the point of research is to understand something well enough that in light of the research, you actually can define it, eg Dark Matter. Or, in a different time, AIDs before we knew what it was.
You can have clusters of related case studies that share the observable effects, and reason and research your way to correlations, and investigate those to discover causation and mechanisms, and infiltrate the "black box" of an unknown thing deeply enough that you account for the whole thing itself.
I think progress on consciousness research in humans is advancing impressively, identifying exactly the kinds of pre and postprocessing done to sensory input and areas of the brain associated with conscious activity and brain to machine interfaces are improving all the time.
Granted the hard problem is still hard and must be respected rather than talked past but the point is we're not stuck. Understanding is gradual and you can model phenomena to the degree that they are understood, closing in from multiple sides.
I hear this a lot but I've never understood why people think it's a deal breaker.
Maybe it's a deal breaker and maybe it's not.
At this point, we're groping in the dark. We don't know enough to stay anything for sure. But we're throwing $billions at it based on the pure hope of somehow getting lucky.
For example, we can't even say for sure that consciousness is something that is isolated in the brain. Neurons exist throughout the human body. Consciousness could very well be a whole body phenomenon --- or not, we really don't know.
Does anyone really know if a human body can be realistically mimicked by software at this point? How much energy and computing power would be required to do so? Is there enough on the planet?
Bottom line: At this point, there are more questions than answers. The one thing we know for sure --- tech billionaires are raking in tons of money from AI.
https://www.hopkinsmedicine.org/health/wellness-and-preventi...
https://www.theguardian.com/business/2025/dec/26/ai-boom-add...
I suppose I don't disagree with the individual observations, but where you lose me is in treating it all like it amounts to some kind of conceptual system crash that precludes progress as a matter of principle. It's not even that I disagree one way or the other so much as it's a matter of having trailed off from the question of the role of definitions in research.
Maybe it helps to consider dark matter. What we have is effectively a placeholder definition based on its observable effects. We don't know if it's WIMPs, axion-like particles, or even some alternative framework for gravity. But we have enough to state meaningful questions about it and iterate toward understanding from a number of directions using a combination of hypothesis, data and experimentation. Finding out what it truly is would be the culmination of research that settles the question rather than something to be stipulated at the start.
So depending on how you look at it, you already have a working definition of consciousness sufficient to organize research, we already have made real progress of the kind that should be impossible if definitions were really dealbreakers, and having "a definition" in the complete sense is something you would never have up until the point the question was settled once and for all, which happens at the end of research rather than the beginning.
I see Wittgenstein mentioned more often in these parts which is awesome, and I think the best Wittgensteinian attitude to adopt here is to turn the tables on this whole question and refuse to agree that there's such a thing as a question of definition that stands between us and research progress.
https://arxiv.org/abs/2412.13145
actual paper.
Because consciousness is not a scientific nor philosophic phenomenon?
Like all references, its symbolic existence exists within the mind of the observer. It's the mistake of reification[1] to forget this authorship and go looking for it as if it exists out there in reality.
I hate to be the bearer of bad news, but the engineer's distaste for soft sciences, and above all sociology, increases the susceptibility of this fallacy. You may be more familiar it's formulation as "the map is not the territory."
Looking for consciousness in the mathematics of ANNs will yield just as much insight as tearing apart the fibers of dollar bills looking for the essence of financial value, distilling paints to find the soul of art, or searching for gender in genetic material. It's a category mistake[2] wrapped in bias. We'd be much more well off learning how the ontology of consciousness is constructed than continuing to apply the hammer of hard science.
1. https://en.wikipedia.org/wiki/Reification_(fallacy)
2. https://en.wikipedia.org/wiki/Category_mistake
[dead]
Eventually it will be extremely difficult to detect in most humans for most humans.
Can we even agree on the definition?