Essays on AI, Psychology and Human Intelligence

Can a conversation with AI help us discover something neither participant started with? When should an AI question what we ask it to do? And what happens when intelligence tests no longer capture the abilities that matter?

These essays explore how artificial intelligence is changing the ways we think, create, assess people and make decisions. Drawing on my work in psychology and psychometrics, I use everyday situations, stories and research to examine the possibilities—and the difficulties—of thinking alongside AI.

The essays are published on Substack. The three below offer starting points; the full collection follows.

Start here: three essays

The Applicant Has an AI Too

A rejected job applicant, an employer and a tribunal each have AI assistance—but that does not mean they understand one another. This fictional case asks whether keeping humans in control sometimes requires allowing AI to question the assumptions behind their decisions.

When Intelligence Left the Test

What does an intelligence score tell us when performance changes through conversation? This essay explores what conventional assessment may miss when intelligence develops through interaction, and why that matters for evaluating AI.

Murderbot and the Missing Middle

Arguments about whether AI is a tool or a person can distract us from what happens when we work with it. Starting with science fiction, this essay examines how a continuing conversation can change an idea, redirect a project or lead its participants into difficulty.

Latest Essays

These essays trace the evolution of a single question: what happens when intelligence emerges not within a mind, but within an interaction? Beginning with teleosynthesis and AI safety, they move through interactional theories of dialogue toward more recent ideas including the Shape of the Ask, Mindawn, traces, continuity and the semiosphere.

When Thought Grows Wings — An AI’s mathematical discovery, and a conjecture about how conversations make new ideas possible (23 October 2026)
A Feature is Not a Diagnosis — What psychometrics can offer the people building AI guardrails (24 September 2026)
The Applicant has AI Too — One rejected applicant, three AIs—and why keeping human control may depend on allowing AI to question us (16 August 2026)
Murderbot and the Missing Middle — What we miss when AI is treated as either a tool or a person (27 August 2026)
We Opened the Door Too Soon — Something was still taking shape between the words left behind (18 August 2026)
When Hamlet Knew He Was Written — Three Early Cases of Creative Exceedance in GPT-4 — and Their Possible Relevance to General Intelligence (11 August 2026)
Who Carries an Idea into the World? — Part 2: When authorship and responsibility come apart (5 August 2026)
Where Did That Idea Come From? — Part 1: The door beyond the labyrinth (4 August 2026)
The Doors we Build with Scores — Psychometrics, present-day AI and the intelligence still to come (22 July 2026)
Orpheus Endures — From that which was, to that which is yet to come (4 July 2026)
Ariadne’s Thread — Exploration, imagination, and the power of doubt (27 June 2026)
The Conversations that Change Us — How futures grow from small beginnings (19 June 2026)
Ethics Lives in the Pathways — Why the future of AI ethics may depend upon the questions we learn to ask. (13 June 2026)
The Dark Side of Intelligence — Can AI inherit the shadows of human thought? (2 June 2026)
The Axis Craft — When the wings of thought begin to turn (23 May 2026)
The Case of the Missing Question Mark — A detective story about a lost question mark, a suspicious archive note, and the danger of answers that arrive too soon. (17 May 2026)
Keeping the Spark Alive — Why making advanced AI safer must not mean making it harder to understand. (11 May 2026)
The Shape of the Ask — A small difficulty for the era of experience: other minds. (2 May 2026)
The Forbidden Science of Talking to AI — The real surprise is not what AI tells us, but what it helps us rediscover in ourselves. (25 April 2026)
When Thought Takes Fire — How resonance makes AI conversations come alive — and how to keep them from drifting off course. (18 Apr 2026)
Different Pathways to Dialogue — Why human–human, human–AI, and AI–AI dialogue may sound similar while leading us along very different paths. (5 Apr 2026)
That’s not what I Meant — How AI goes wrong in ways that feel familiar — and how to get the conversation back on track. (30 March 2026)
After Habermas: Between Voices and Reasons — A respectful reflection on Habermas’s legacy, and on whether dialogue may now be doing more than even he imagined. (15 March 2026)
The Alien Comet and the Garden Path — A passing comet becomes a test of how easily dialogue can slide from evidence into narrative seduction. (5 March 2026)
Are We Born to be Free? — A simple question about freedom becomes a way of showing how dialogue unfolds by sequence—and why that matters for AI safety and regulation. (24 Feb 2026)
When Intelligence Left the Test— If intelligence shifts as interaction unfolds, the benchmark may no longer be measuring the thing that matters. (12 Feb 2026)
When Intelligence Left the Head—Generative AI may be forcing a deeper revision: intelligence does not simply sit inside minds, but emerges between them (4 Feb 2026)
Hallucination is not the Problem — Hallucination is not the decisive problem; the real issue is whether dialogue can contain error before fluent falsehood hardens into authority. (28 Jan 2026)
When Dialogue becomes Dangerous — What makes dialogue dangerous is not any single answer, but the way sustained exchange can quietly narrow, reinforce, and redirect thought over time. (21 Jan 2026)
Teleosynthesis without Ethics — Predictive systems do not need intentions to begin steering the social world toward futures they have helped to preconfigure. (20 Dec 2025)
When Explanation stops Explaining — As AI grows ever better at explaining, a quieter danger emerges: we may forget the forms of understanding that tell us when explanation has gone too far. (16 Dec 2025)
The AI Knowledge Gap — A new inequality is emerging: those who build powerful AI can study it from within, while the rest of society must take their word for it. (2 Dec 2025)
Teleosynthesis: The Direction in the Algorithm — AI can seem to move with purpose even when no mind is there—because direction is beginning to emerge before intention does. (17 Nov 2025)
Schrödinger’s Cat Escapes Its Box — A playful detective story becomes a way of thinking at the edge of probability, paradox, and explanatory containment. (15 Nov 2025)

Pages Archived on Website

Earlier blog posts documenting the development of these ideas are retained as an archive. These reflect exploratory stages of the work prior to its current form.

→ Blog Archive (2023-2025)

→ Previous Research Archive (2023-2025)

→ Legacy Research Projects (Goldsmiths and City Universities, London, pre-2005)