Kris Van der Bijl

Arthur Goldstuck’s latest book, The Hitchhiker’s Guide to AI: A Handbook for All, opens the world of this ‘new’ technology to the lay reader. It’s highly accessible to those – like me – who pretend to know about technology and what it might mean for our present and future. When artificial intelligence (AI) and Machine Learning (ML) appeared to explode, as ChatGPT was launched in late 2022, backlash from creators coalesced around issues of plagiarism and how to account for where these technologies were getting their information from. Educators were interested in what this means for the future of education. I grew interested in what these new and ever-coming technologies mean when we say authority.

Goldstuck writes his text in such a way that trust in his authority on the topic remains tenuous. He foregrounds himself as an expert in technology, and the text even has a preface written by an AI industry expert. Goldstuck’s CV as a tech journalist also adds to his implicit credentials. And it would be bizarre for Goldstuck to praise AI’s functionality and not use it in his text. There’s a good thought experiment to be had about authority and authorship here. Whether or not he used AI is of little consequence. What constitutes authority in writing has been, and likely will remain, challenging to define, regardless of new technology.

Goldstuck acknowledges early on that “I used generative AI to help write this book. But the emphasis is not help, not write” (16). The emphasis on ‘help’ is essential when thinking about authority in writing. It is a myth that writers write alone. What they write is rarely the final product. There are editors, lawyers, friends-cum-proofreaders, sensitivity readers, and the like. Anyone whose advice and knowledge are sought for the final product to be as best as possible helps the text become more credible, believable, and worth reading.

These inputs are often unseen in the final product, and why would it be? Critics are taught to scrutinize writers and their words. A book showing an editor’s split infinitive corrections will do a writer no favour. Mistakes are corrected before publication because they allow critics to poke holes in a piece of work they can use to attack an argument. For writers, the more bungs, the better. Industry standards require no mistakes. If we believe a book without faults, we can view it as an authority on their respective topic.

I use the word ‘authority’ with little clarity. I am using it to connote a personal interest in how an author makes us believe what they say is correct and truthful. It’s how they make us nod along. One technique that Goldstuck uses is to refer to events – like conferences – that he attended on AI and ML while writing the book. It’s a sign of research, something that we readers can use to confirm that he knows a particular topic. This tactic works alongside the unknowable edits and corrections so that we trust him. It amalgamates to grant him authority. 

That a text is proofread and edited by a collection of nameless people is not new. Even Homer is seen as a collection of spoken word poets coalesced into a single author. But what is new is Goldstuck’s never-explicit use of AI technology. Almost all word processors come with automated correction software. Phrases like “lacks basic editing” refer to the writer’s lack of spellcheck software, a missing link that affects how their writing is perceived. For better or worse, appearing clean is essential for humans and books. The difference is that the cleaning and fixing done by AI is like a new infectious disease. Its general acceptance as something that does not need to be disclosed remains in a future decision. We might still need information about it.

Goldstuck’s text functions to help us understand AI and ML technology so that we can live better lives. Businessmen will find explanations on how to automate production. Goldstuck even gives examples of which tools are currently on the market and what their pros and cons are. He talks about AI and ML’s medical, banking, and education benefits. No segment of society appears to be left out. Goldstuck proves himself to be a master of the AI and ML fields because the text is informative. Even an English literature student could follow along and pretend to know what the technology means. Technology’s function is to make the pity and sorrow of human life more bearable. It’s a natural progression from creating communities to ensure a human does not die alone in the wilderness. We help each other, and we develop tools to help us help. 

Creatives are not left in the lurch about their benefits. Goldstuck is a creator like them. The developing art of writing prompts for AI software is given an entire chapter. It is an industry that creatives might already have a head-start in. I picture “AI Prompt Writer” added alongside “SEO Expert” and “Social Media Manager” on new Journalism graduates’ LinkedIn pages. My image lets me sneak back to questions on authority. I can feign SEO knowledge through a bought diploma, but most companies will see the knowledge and ability that I lack. The ability to write good prompts might also be a case of sifting out the bad to see what is left. If I were desperate for employment, then what stops me from using a prompt to write a prompt? If I were smart, I could write an AI prompt generator and start an industry of prompt-for-prompt writing software. I now imagine prompt-for-prompt-for-prompt-(n+1) devices all to write copy for a hand cream. Of course, the consumer will need to trust the copy.

The latter half of the text addresses plagiarism concerns. I’m overstating for emphasis by saying AI and ML models don’t say where they get their information. They also don’t explicitly state what they do and do not make up. Like wannabe freelancers, they tend to over-emphasize their knowledge base to land gigs they might need to be more competent to complete. But if the consumer is OK with the product, who cares? Goldstuck does reference intermittently, yet we can also call out his mistakes as an author. The polishing by numerous unknown faces does this before publication. He possesses authority through an unspoken pact about a physical book’s trustworthiness. We know not to trust technology because we’ve all seen it fail. Authors of physical books don’t face the same level of scrutiny because of discrete industry standards. 

A misinformed published text is a loss of revenue, public faith, and a waste of paper, while a misinformed website is a symptom of the internet. 

I’m using these broad generalizations to talk about the peculiar brilliance of Arthur Goldstuck’s ability to be the author of The Hitchhiker’s Guide to AI. What’s striking is that a physical book can be used as a medium to inform the public about a topic that it is distinct from. The authorship process of how to trust what is written is central to the text. It informs the reader not to fear new AI and ML technology at a content level. It also does it through the rhetoric of authorship, even though a history of convention of industry standards hides its processes. If we can’t trust the authority of Goldstuck, why not trust Google? It’s not like Big Tech has let us down before. 

Book Title:                              The Hitchhiker’s Guide to AI: A Handbook for All

Author:                                    Arthur Goldstuck

Publisher:                               Pan Macmillan South Africa

Place of Publication:              Johannesburg

Year of Publication:               2024

Pages:                                    215

ISBN:                                      978-1-77010-896-7

Publisher Website:                 www.panmacmillan.co.za