Review: Feeding the AI Machine

(James Muldoon, Mark Graham and Callum Cant, Canongate, £20)

Captivated as we are by the wonder of seeing Al apps summarise long reports and produce schemes of work in a matter of seconds, it perhaps hasn't occurred to enough us how all the 'magic' involved actually works - that the data Al relies on has to first be tagged and processed before it can be used.

This book examines the working lives of the people involved in precisely that. Alongside some harrowing workplace stories and testimonies, the authors present some suggestions as to how things might be improved.

The detailed research and the issues highlighted herein are a timely corrective to the techno-optimism of Al advocates, though be advised that the authors make little secret of their political leanings. The book's content could be a good springboard for classroom discussion, but approach with care...

This book was first reviewed in Teach Secondary magazine.

The published review (above) differed from the one I actually submitted. My original one is slightly different. In case you’re interested, here it is:

My original review

We are probably all captivated by the wonder of seeing AI apps producing a scheme of work in almost no seconds flat, or summarising the contents of a long report. Perhaps, though, it hasn’t occurred to us that behind all this ‘magic’, the data that AI relies on had to be tagged and processed in order that it could be used.

This book looks at the working lives of a range of people involved. Some of the stories are harrowing, but the authors present some ideas about how things might be improved. One or two of these might work, but most probably wouldn’t.

The research and issues highlighted provide a good antidote to the techno-optimism sometimes associated with AI, but the book is, frankly, spoilt in my opinion by the not-so-well-hidden left-wing bias. If you believe that politics has no place in the classroom, approach this book with care.