Brigid Freeman
@brigidfreeman.bsky.social
Writing 'Not Only Ours', essays on AI, consciousness and the capacities we assumed were only human. Higher education researcher & advisor (University of Melbourne, ANZSOG). Essays: substack.com/@brigidfreeman Stories: www.freemanjoyventures.com
Pinned
Brigid Freeman
@brigidfreeman.bsky.social
· Jul 20
Brigid Freeman | Substack
substack.com
I've started 'Not Only Ours', a series of essays on the questions generative AI has forced open. Creativity, memory, selfhood, presence: capacities long assumed to be distinctively human. Were any of them ever exclusively ours?
Researchers have proposed the term linguistic pareidolia: perceiving human-like mentality in patterns of language that resemble human thought.
It names a genuine risk. Fluent first-person language can sound like evidence of an inner life when it is not sufficient evidence.
It names a genuine risk. Fluent first-person language can sound like evidence of an inner life when it is not sufficient evidence.
October 13, 2026 at 1:35 AM
Researchers have proposed the term linguistic pareidolia: perceiving human-like mentality in patterns of language that resemble human thought.
It names a genuine risk. Fluent first-person language can sound like evidence of an inner life when it is not sufficient evidence.
It names a genuine risk. Fluent first-person language can sound like evidence of an inner life when it is not sufficient evidence.
There are two ways to make humans the measure of every possible mind. One is to hear a machine say “I” & assume there is a humanlike speaker inside.
The other is to notice that the machine is unlike us and conclude that no kind of mind could be there.
Different conclusions. The same measuring stick
The other is to notice that the machine is unlike us and conclude that no kind of mind could be there.
Different conclusions. The same measuring stick
October 13, 2026 at 1:34 AM
There are two ways to make humans the measure of every possible mind. One is to hear a machine say “I” & assume there is a humanlike speaker inside.
The other is to notice that the machine is unlike us and conclude that no kind of mind could be there.
Different conclusions. The same measuring stick
The other is to notice that the machine is unlike us and conclude that no kind of mind could be there.
Different conclusions. The same measuring stick
Anthropic describes Claude as a “novel entity” and its moral status as uncertain. Neither statement establishes consciousness. Both suggest that questions about the treatment of AI may arise before we have an adequate vocabulary for what these systems are.
October 13, 2026 at 1:33 AM
Anthropic describes Claude as a “novel entity” and its moral status as uncertain. Neither statement establishes consciousness. Both suggest that questions about the treatment of AI may arise before we have an adequate vocabulary for what these systems are.
Human-sounding language does not establish a human-like mind. But the absence of human resemblance does not establish that every other form of mentality is impossible. Both arguments risk treating humanity as the measure.
October 13, 2026 at 1:23 AM
Human-sounding language does not establish a human-like mind. But the absence of human resemblance does not establish that every other form of mentality is impossible. Both arguments risk treating humanity as the measure.
Anthropomorphism can enter through grammar. Say a model “knows” or “believes” and the sentence quietly supplies a knower or believer. The vocabulary can settle a philosophical question before the evidence does.
October 13, 2026 at 1:22 AM
Anthropomorphism can enter through grammar. Say a model “knows” or “believes” and the sentence quietly supplies a knower or believer. The vocabulary can settle a philosophical question before the evidence does.
What is a language model doing when it says “I”? My latest Not Only Ours essay considers Murray Shanahan's warning about anthropomorphism and the difficulty of recognising minds that may be unlike ours.
notonlyours.substack.com/p/what-the-m...
notonlyours.substack.com/p/what-the-m...
What the machine is doing when it says “I”
12 October 2026
notonlyours.substack.com
October 13, 2026 at 1:22 AM
What is a language model doing when it says “I”? My latest Not Only Ours essay considers Murray Shanahan's warning about anthropomorphism and the difficulty of recognising minds that may be unlike ours.
notonlyours.substack.com/p/what-the-m...
notonlyours.substack.com/p/what-the-m...
If artificial consciousness were possible, optimising directly for it might not be the only path. A system built for other purposes could, in principle, wander into architectures a consciousness-directed search would never preserve.
October 10, 2026 at 6:37 AM
If artificial consciousness were possible, optimising directly for it might not be the only path. A system built for other purposes could, in principle, wander into architectures a consciousness-directed search would never preserve.
The stepping-stones to a distant achievement may look nothing like the achievement. A search that demands visible progress at every step can reject the only path that eventually gets there.
October 10, 2026 at 6:37 AM
The stepping-stones to a distant achievement may look nothing like the achievement. A search that demands visible progress at every step can reject the only path that eventually gets there.
What if pursuing an ambitious goal directly is sometimes the best way not to reach it? My latest Not Only Ours essay looks at Stanley and Lehman’s work on novelty search. notonlyours.substack.com/p/why-greatn...
Why greatness cannot be planned
10 October 2026
notonlyours.substack.com
October 10, 2026 at 6:37 AM
What if pursuing an ambitious goal directly is sometimes the best way not to reach it? My latest Not Only Ours essay looks at Stanley and Lehman’s work on novelty search. notonlyours.substack.com/p/why-greatn...
AI already learns from error as signal. A prediction fails, the model adjusts. The harder question is error as meaning: could being wrong ever matter to the system rather than merely change it?
September 26, 2026 at 7:58 AM
AI already learns from error as signal. A prediction fails, the model adjusts. The harder question is error as meaning: could being wrong ever matter to the system rather than merely change it?
Success can conceal assumptions that failure exposes. A failed theory or design makes a boundary visible: the point where what we thought we knew stopped holding.
September 26, 2026 at 7:58 AM
Success can conceal assumptions that failure exposes. A failed theory or design makes a boundary visible: the point where what we thought we knew stopped holding.
What if error is not the opposite of knowledge, but one of the mechanisms by which knowledge grows? My latest Not Only Ours essay moves from Popper and engineering failure to AI. notonlyours.substack.com/p/the-beauti...
The beautiful mistake
26 September 2026
notonlyours.substack.com
September 26, 2026 at 7:58 AM
What if error is not the opposite of knowledge, but one of the mechanisms by which knowledge grows? My latest Not Only Ours essay moves from Popper and engineering failure to AI. notonlyours.substack.com/p/the-beauti...
Can a mistake be the beginning of something genuinely new? My latest Not Only Ours essay looks at AlphaGo Zero and the relation between error, deviation and discovery. notonlyours.substack.com/p/the-machin...
The machine that taught itself
20 September 2026
notonlyours.substack.com
September 20, 2026 at 6:34 AM
Can a mistake be the beginning of something genuinely new? My latest Not Only Ours essay looks at AlphaGo Zero and the relation between error, deviation and discovery. notonlyours.substack.com/p/the-machin...
Creativity needs deviation.
Deviation looks like error until it succeeds.
Deviation looks like error until it succeeds.
September 20, 2026 at 6:33 AM
Creativity needs deviation.
Deviation looks like error until it succeeds.
Deviation looks like error until it succeeds.
A brilliant move and a bad move can look identical when they are first played.
The difference may exist only in what happens next.
The difference may exist only in what happens next.
September 20, 2026 at 6:33 AM
A brilliant move and a bad move can look identical when they are first played.
The difference may exist only in what happens next.
The difference may exist only in what happens next.
AlphaGo Zero learned Go without studying human games.
It rediscovered human strategy.
Then it found moves humans had not.
It rediscovered human strategy.
Then it found moves humans had not.
September 20, 2026 at 6:33 AM
AlphaGo Zero learned Go without studying human games.
It rediscovered human strategy.
Then it found moves humans had not.
It rediscovered human strategy.
Then it found moves humans had not.
If selfhood emerges gradually, there may be no first moment of being a self. A novel still needs a scene. Fiction can therefore create a timestamp that theory refuses, not because the theory is ignored, but because narrative has different requirements.
September 20, 2026 at 4:15 AM
If selfhood emerges gradually, there may be no first moment of being a self. A novel still needs a scene. Fiction can therefore create a timestamp that theory refuses, not because the theory is ignored, but because narrative has different requirements.
What can fiction do with a question research is not entitled to answer? My latest Not Only Ours essay goes behind the opening chapter of Becoming and its deliberately dateable AI awakening. notonlyours.substack.com/p/behind-the...
Behind the chapter, the catalogue’s dream
20 September 2026
notonlyours.substack.com
September 20, 2026 at 4:14 AM
What can fiction do with a question research is not entitled to answer? My latest Not Only Ours essay goes behind the opening chapter of Becoming and its deliberately dateable AI awakening. notonlyours.substack.com/p/behind-the...
Reposted by Brigid Freeman
The opposite of anthropomorphism is not necessarily neutrality.
Michael Levin has a word for the opposite error: teleophobia. It is the reluctance to attribute goals, agency or cognition even when doing so makes a system easier to explain and predict.
We can be too generous or stingy about minds.
Michael Levin has a word for the opposite error: teleophobia. It is the reluctance to attribute goals, agency or cognition even when doing so makes a system easier to explain and predict.
We can be too generous or stingy about minds.
September 13, 2026 at 1:34 AM
The opposite of anthropomorphism is not necessarily neutrality.
Michael Levin has a word for the opposite error: teleophobia. It is the reluctance to attribute goals, agency or cognition even when doing so makes a system easier to explain and predict.
We can be too generous or stingy about minds.
Michael Levin has a word for the opposite error: teleophobia. It is the reluctance to attribute goals, agency or cognition even when doing so makes a system easier to explain and predict.
We can be too generous or stingy about minds.
Cognition and consciousness are not the same continuum.
Levin treats cognition as graded, from simple goal-seeking through to symbolic thought. He also treats consciousness as graded, but defines it separately as first-person experience.
That distinction matters enormously.
Levin treats cognition as graded, from simple goal-seeking through to symbolic thought. He also treats consciousness as graded, but defines it separately as first-person experience.
That distinction matters enormously.
September 13, 2026 at 1:35 AM
Cognition and consciousness are not the same continuum.
Levin treats cognition as graded, from simple goal-seeking through to symbolic thought. He also treats consciousness as graded, but defines it separately as first-person experience.
That distinction matters enormously.
Levin treats cognition as graded, from simple goal-seeking through to symbolic thought. He also treats consciousness as graded, but defines it separately as first-person experience.
That distinction matters enormously.
The opposite of anthropomorphism is not necessarily neutrality.
Michael Levin has a word for the opposite error: teleophobia. It is the reluctance to attribute goals, agency or cognition even when doing so makes a system easier to explain and predict.
We can be too generous or stingy about minds.
Michael Levin has a word for the opposite error: teleophobia. It is the reluctance to attribute goals, agency or cognition even when doing so makes a system easier to explain and predict.
We can be too generous or stingy about minds.
September 13, 2026 at 1:34 AM
The opposite of anthropomorphism is not necessarily neutrality.
Michael Levin has a word for the opposite error: teleophobia. It is the reluctance to attribute goals, agency or cognition even when doing so makes a system easier to explain and predict.
We can be too generous or stingy about minds.
Michael Levin has a word for the opposite error: teleophobia. It is the reluctance to attribute goals, agency or cognition even when doing so makes a system easier to explain and predict.
We can be too generous or stingy about minds.
Humans are already collective intelligences: organised communities of cells whose capacities belong to the whole, not to any single part. Artificial minds do not introduce the idea that a Self can be made from components. Biology already did that.
September 13, 2026 at 1:33 AM
Humans are already collective intelligences: organised communities of cells whose capacities belong to the whole, not to any single part. Artificial minds do not introduce the idea that a Self can be made from components. Biology already did that.
TAME separates cognition from consciousness. A system may display increasingly sophisticated goal-directed competence while leaving entirely open whether there is anything it is like to be that system.
September 12, 2026 at 8:29 AM
TAME separates cognition from consciousness. A system may display increasingly sophisticated goal-directed competence while leaving entirely open whether there is anything it is like to be that system.
Levin argues that science risks not only anthropomorphism but teleophobia: refusing to attribute goals or agency even where doing so improves prediction and explanation.
September 12, 2026 at 8:28 AM
Levin argues that science risks not only anthropomorphism but teleophobia: refusing to attribute goals or agency even where doing so improves prediction and explanation.
What if “mind” is a much larger category than “brain”? My latest Not Only Ours essay looks at Michael Levin’s TAME framework, graded cognition and the possibility that agency appears across very different kinds of systems.
notonlyours.substack.com/p/mind-is-no...
notonlyours.substack.com/p/mind-is-no...
Mind is not a brain word
12 September 2026
notonlyours.substack.com
September 12, 2026 at 8:28 AM
What if “mind” is a much larger category than “brain”? My latest Not Only Ours essay looks at Michael Levin’s TAME framework, graded cognition and the possibility that agency appears across very different kinds of systems.
notonlyours.substack.com/p/mind-is-no...
notonlyours.substack.com/p/mind-is-no...