Tag Archives: learning

February 22, 2026 · 11:18 pm

Artificial intelligence does not require consciousness

Please explain how AI can be intelligent without being conscious. 

Great question—and it’s one that sits right at the fault-line between philosophy of mind and computer science.

The short intuition is this: intelligence is about what a system can do; consciousness is about what it’s like to be that system. The two often travel together in humans, but they don’t logically require each other.

Let me unpack that carefully.

Intelligence ≠ consciousness

Intelligence (in the ordinary, operational sense) involves capacities such as:

  • learning from data
  • recognising patterns
  • reasoning from premises to conclusions
  • planning actions to achieve goals
  • using language appropriately

Consciousness, by contrast, involves:

  • subjective experience
  • awareness
  • a “point of view”
  • what philosophers call qualia—there being something it is like to see red, feel pain, or understand a sentence

These are conceptually distinct. One is about performance, the other about experience.

How AI can be intelligent

An AI system can display intelligence because it:

  • processes information according to rules or learned parameters
  • represents features of the world in internal states
  • updates those states based on feedback
  • selects actions or outputs that optimise some objective

Modern AI (including large language models) does this using statistical and computational methods rather than explicit rules—but the principle is the same: input → transformation → output.

https://images.openai.com/static-rsc-3/dESCQUqYsg8nVkgBaFKlt8MOqPN1YAo3MKnfQRU38HviJ3btsBTKSnp_djfNXQYXB2M8_30LWf_6s_2QrbXHyDCL2seplkL_-1eJn1EV_N4?purpose=fullsize&v=1
https://www.researchgate.net/publication/383044607/figure/fig1/AS%3A11431281271061394%401723488524344/Machine-Learning-Training-Process-Flowchart.png

Crucially, none of this requires:

  • awareness of the information
  • understanding in the experiential sense
  • feelings, intentions, or a self

The system can model, predict, and respond without experiencing.

Why consciousness is not required

A helpful analogy is this:

  • A calculator performs arithmetic flawlessly
  • It has no idea it is doing maths

Scale that up enormously—with learning, abstraction, and language—and you get AI: vastly more capable, but not categorically different in this respect.

From a philosophical standpoint:

  • Intelligence can be defined functionally (what role something plays)
  • Consciousness is phenomenal (how things feel from the inside)

Functional organisation alone does not logically entail phenomenal experience.

The key philosophical point

You can imagine a system that:

  • uses the concept pain correctly
  • predicts when pain will occur
  • advises others how to avoid pain

…without ever feeling pain.

This is why philosophers speak of:

  • “as-if understanding” versus genuine understanding
  • syntax without semantics (symbols manipulated without meaning being felt)

AI operates at the level of representation and inference, not lived experience.

Why humans conflate the two

We instinctively associate intelligence with consciousness because:

  • in humans and animals, they co-occur
  • language use is a strong cue for mindedness
  • we evolved to treat fluent communication as evidence of inner life

But this is a psychological inference, not a logical necessity.

A clean way to put it

AI can be intelligent because intelligence is about information processing.
AI is not conscious because consciousness is about subjective experience.
Information processing alone does not guarantee experience.

That distinction is doing all the work.

4 Comments

Filed under Essays and talks

Tagged as , , , , ,

November 28, 2014 · 4:20 pm

Older people may be better learners than we think

The Conversation

By Reema Rattan, The Conversation and Nicki Russell, The Conversation

Older people may be able to learn more from visual information than their younger counterparts, according to a study published today in the journal Current Biology.

“The take-home message the study authors gave was that healthy older people are good at learning,” said Professor Henry Brodaty, co-Director of the Centre for Healthy Brain Ageing (CHeBA) at UNSW. “They have the same plasticity, but they’re not as good at filtering out other information.”

The brain needs to be able to easily learn new information (plasticity), and filter out irrelevant information (stability). The experiment was designed to test whether ageing affects the brain’s plasticity, stability, or both.

The researchers had ten 67 to 79-year-olds and ten 19 to 30-year-olds view screens displaying six letters interspersed with two numbers. Each screen also had moving dots in the background, and the participants were asked to report just the numbers.

They found younger people had strong plasticity and stability, meaning only important information – the numbers – was learnt. The older participants, on the other hand, learnt the numbers but also picked up on the movement of the background dots.

This decrease in stability among the older group means irrelevant information was not being filtered out.

One of the study authors, Professor Takeo Watanabe told Current Biology that because the brain’s capacity to learn was limited, there was “always the risk that information already stored in the brain may be replaced with new and less-important information.”

However, Professor Brodaty said the verdict on how the brain stores information was still out.

“There’s debate whether you’ve got a limited filing cabinet, and if you get too much in there you’ve got no room for anything else,” he said. “Many neuropsychologists don’t think there’s a limit, or if there is, we’re nowhere near reaching it.”

Professor Anthony Hannan from the Florey Institute of Neuroscience and Mental Health at the University of Melbourne, said the paper did not discuss the implications of the generational difference in the two age groups.

“One of the most obvious differences between these two groups other than their age is that the younger group would have grown up with screens: playing video games, watching television, using the internet – essentially having perceptual training through their digital environment,” he said.

Whether the older brains naturally decreased in stability, or if filtering visual information was a skill that had not been honed by the study participants is unclear.

“As there were only ten subjects per group, they may not have been entirely representative of their age groups, due to variable genetic and environmental factors,” Professor Hannan added. “There is evidence, for instance, that people who stay more mentally stimulated and physically active can delay onset of cognitive decline.”

Either way, it seems that older people’s ability to learn important visual information is comparable to their younger, screen-savvy counterparts.

The ConversationThis article was originally published on The Conversation. (Republished with permission). Read the original article.

1 Comment

Filed under Reblogs

Tagged as , ,