Rachel Hamilton – Author and Academic at MAC 2026
Can Machines Make Us Feel?
Why do our brains insist on finding a mind where none exists - and what happens when that illusion is used against us?
Rachel Hamilton is a writer and academic whose work sits at the intersection of AI, literature, and human emotion. In her MAC session, she explores the fundamental gap at the heart of human-AI interaction: the space between the genuine feeling an AI produces in us and the total absence of feeling on the other side.
Drawing on the "ELIZA effect" and original research into collaborative play, Rachel investigates why we respond to language models as if someone is actually there. From "grieving" when ChatGPT 4 is updated to analysing the specific moments where AI language collapses, she reveals the evolutionary mechanisms that make us vulnerable to machine-generated empathy - and why simple disclosure may never be enough to break the spell.
What you will learn in this talk:
1 1
STEP 1
The ChatGPT Grief:
Why users experience genuine loss when a software update changes an AI’s "personality".
2 2
STEP 2
The Poetry Paradox:
Why readers typically fail to identify the human author in a blind test between a machine and a Pulitzer Prize winner.
3 3
STEP 3
The ELIZA Effect:
How our brains are hard-wired to find "minds" in simple code, regardless of whether we know it is software.
4 4
STEP 1
The Collapse of the "I":
The specific linguistic threshold where an AI’s ability to describe the human world fails and its lack of an inner life is revealed.
5 5
STEP 2
The Empathy Trap:
Why the same triggers that make a poem moving are being optimized at scale for phishing scams and political messaging.
Rachel Hamilton Bio
Rachel Hamilton is an award-winning children’s author and a PhD researcher in Creative Writing at the University of Bristol. Her research focuses on human-AI creative collaboration and the ethics of machine-generated empathy. The author of six novels for young people, Rachel also teaches at Bath Spa University, where she champions AI literacy in the classroom. Before her career in academia, she worked in advertising, stand-up comedy, and prison administration.
Rachel Hamilton Bio
Would you say there is a value difference between generating an idea from an AI that was expanded upon by a human, and an idea that came from a human that they used AI to turn into a workable story?
I suppose some people would argue that the idea is the valuable thing and that expanding it is just a craft exercise. But, as Christopher Booker famously argued, there are only seven basic plots, and it’s not unusual for authors to have external input into their ideas – from agents, publishers, editors, tutors etc. So, while some ideas are definitely more powerful than others, I tend to think that creative value comes less from a unique premise than from what an author does with it. In the ‘AI-idea – human expansion’ model, while an AI may have provided the spark, the human will have made thousands of decisions and choices about what to include, what to leave out, what to focus on, and so on. Those hours they’ve spent on it will have injected their own human perspective, lived experience, and emotional stakes. Whereas, in the ‘human-idea – AI expansion model’, while the idea may have been an expression of human experience, the AI will have polished and built it through prediction and probability patterns rather than through anyone’s lived truth, potentially flattening human quirks and contradictions into a statistically optimised average.
So, my personal feeling is that a great concept isn’t that hard to find, and it’s the human development that gives stories their value. But I’m happy to be challenged on that!
Is the purpose of reading literature to engage remotely with another mind (in some sense) regardless of the story and characters – without analysis/mindfulness we can imagine that we might not always pick up when a thing is AI writing – as in your AI phone money scam example – so if future AI could generate literature that no-one can identify as AI would there be any point in reading it – at the deeper level?
This is something I’m really interested in – why do we read literature? At a superficial level, we could say it’s simply to receive a story. Perhaps we’re reading for escape or entertainment, and, if so, then undetectable AI literature would arguably have value and work as effectively as human literature. But I believe, along with many others, that a key reason why we read is to establish some kind of existential connection – to engage with another consciousness and explore what it would feel like to be someone who notices these particular things in this particular order. In this way, we’re treating literature as a window into other minds.
If you support that view, then I guess the answer to your question depends on how we perceive AI and LLM output. Currently, you could argue that reading undetectable AI literature would be quite a hollow experience, in that you’d be trying to empathise with a mind that isn’t there. You’d be engaging with a statistical machine trained to pattern-match what humans have created before – a mirror rather than a window. But if finding value in literature is partly a construction by the reader, then we have always projected onto texts even when we know little about the author, and some of the most renowned writers, like Homer or Shakespeare, may be composite or may have worked extensively with collaborators. Also, as AI becomes increasingly sophisticated, perhaps it would be interesting to have a window into the AI mind and engage with a different kind of consciousness. But you’d want transparency. You’d want to know that it was an AI mind you were engaging with.
I guess you could make an analogy with visiting a place vs. visiting a perfect simulation of it. The experiences might be hard to distinguish, but what you’d done would be different.
