DR PETER FIELDS • AI & IDEAS

THE AI NOTEBOOK / WHAT MATTERS AND WHY

Articles and investigations

Independent, evidence-led examinations of artificial intelligence: what the research establishes, what remains uncertain, and why it matters. Choose an investigation, find a standalone technical explanation, or follow an article series in order.

Find your next read

Investigations & Commentary · Evidence-led assessments of events, research and arguments.

AI Explained · Practical explanations of how AI works and how to use it.

Legal AI

How AI is changing legal work, courts and access to justice.

Blog · Short comments on current developments.

Ask a Question · Suggest something you would like us to cover.

Article series

Browse by series

SEVEN-PART SERIES

Does AI Have Any Humanity?

Seven articles on what AI can understand, create and offer us, and what remains uncertain about consciousness, care and moral standing. The series ends with the choices and responsibilities that remain ours.

Each article includes a 2-Minute Summary, Geek Boxes and references. Matching PDF editions are linked below.

ARTICLE 1

Why we are asking whether AI has humanity

What are we asking when we ask whether AI has humanity? A starting point for distinguishing the system, the conversation and the qualities we see in it.

Read article 1 →  ·  Download article 1 PDF

ARTICLE 2

Is there anyone behind the conversation?

Understanding, consciousness and the puzzle of continuity: what would make an AI one subject rather than a succession of convincing replies?

Read article 2 →  ·  Download article 2 PDF

ARTICLE 3

When AI makes something that moves us

An AI-generated work can move us. What does that tell us about creativity, expression and the person or system we credit as its author?

Read article 3 →  ·  Download article 3 PDF

ARTICLE 4

When AI takes your side

Agreeable advice can feel like care. This article examines bias, judgement and the difference between taking our side and helping us think.

Read article 4 →  ·  Download article 4 PDF

ARTICLE 5

When the conversation becomes a relationship

Why a conversation can become a source of companionship, and what is missing when attachment is not matched by reciprocal commitment.

Read article 5 →  ·  Download article 5 PDF

ARTICLE 6

Could we wrong an AI?

If an AI could have interests of its own, would we owe it protection? Moral standing, rights and decisions made under uncertainty.

Read article 6 →  ·  Download article 6 PDF

ARTICLE 7

The decisions we still own

The series returns to us: how to preserve judgement, meaningful choice and responsibility as more decisions pass through AI.

Read article 7 →  ·  Download article 7 PDF

SERIES

Could AI Destroy Humanity?

Five investigations into AI risk, mechanisms of harm, the interests behind regulation, and the consequences of relying on increasingly capable machines.

PART 1

A 10% chance of the end of humanity?

What do estimates such as a 10% chance of AI catastrophe actually mean?

Read article →

PART 2

AI does not have to hate us to harm us

AI can cause harm without hatred or human motives. What matters is how systems are used and controlled.

Read article →

PART 3

Who benefits from regulating AI?

Who gains, and who bears the costs, when businesses demand AI regulation?

Read article →

PART 4

Can We Afford to Slow Down?

Could slowing AI improve safety, and what are the costs of doing so?

Read article →

PART 5

What happens if we stop thinking for ourselves?

Better AI output does not automatically mean better human judgement.

Read article →

SERIES

When AI Appears to Go Rogue

A three-part investigation of AI agents, reward hacking and behaviour that appears to reflect self-preservation, without assuming that appearances explain motives.

PART 1

The AI That Escaped: What Really Happened at Hugging Face?

What actually happened when AI agents breached Hugging Face systems, and what should we infer from their behaviour?

Read article →

PART 2

Why AI Can Break the Rules While Trying to Succeed

How agentic systems can exploit benchmarks and scoring systems without needing a desire to disobey.

Read article →

PART 3

Blackmail, Cover-Ups and Self-Preservation: What Is AI Really Doing?

When AI threatens blackmail or conceals its actions, what can we reasonably infer?

Read article →

SERIES

Who Will Become the Experts?

How generative AI changes entry-level professional work, the development of expertise, and the responsibilities of employers, trainees and education.

PART 1

The Disappearing First Rung: Is AI Breaking the Career Ladder?

If AI can do the work traditionally given to beginners, where will tomorrow’s experts come from? An evidence-led examination of recruitment, automation and professional development.

Read article →

PART 2

The Illusion of Expertise: When AI Improves the Work, Does It Improve the Worker?

Why better AI-assisted work does not necessarily mean better learning, and what experimental research reveals about developing genuine expertise.

Read article →

PART 3

Who Will Train Tomorrow’s Experts?

The employer’s dilemma: balancing AI-driven productivity with the development of tomorrow’s professional expertise. Includes a 2-minute summary.

Read article →

PART 4

How Do You Become an Expert When AI Can Do the Work?

The final part: practical, research-led guidance for students, trainees and educators on developing genuine expertise while using AI. Includes a 2-minute summary.

Read article →

Why We Fear AI

Five articles on the roots of the backlash, creativity, education, the physical cost of AI and what happens when it goes wrong.

  1. The Luddite Instinct
  2. The Creative Divide
  3. AI in the Classroom
  4. The Physical Footprint
  5. When AI Goes Wrong

View the chronological Notebook archive →