Futurism Discovers Cliff’s Notes and Calls It the Death of Reading

,

Futurism has published a worried little dispatch about an Axios publisher using AI to help him “read” The Last of the Mohicans. The horror, apparently, is that he used a machine to remember where he was in a difficult old novel, answer a question, and help him proceed.

One imagines Futurism discovering a bookmark next.

The piece’s underlying claim is that AI has created a new form of intellectual abdication: instead of reading, people now ask a machine to do the thinking for them. It is a tidy argument, emotionally satisfying, and about as historically informed as announcing that social media invented gossip.

People have always used intermediaries to help them read, avoid reading, remember what they read, or pretend to have read. We have had Cliff’s Notes, Coles Notes, Monarch Notes, SparkNotes, abstracts, annotated editions, lecture notes, book reviews, study guides, executive briefings, encyclopedias, digests, newspaper explainers, and that venerable professional institution: somebody else’s research assistant.

The book-summary industry was not subtle about its purpose. Capitol Reader, which sold summaries of political books, including a summary of my 2005 book Outfoxed: Rupert Murdoch’s War on Journalism, advertised its service this way: “We’ve done the reading so you don’t have to.”

That is not a sinister chatbot’s hidden agenda. That is the marketing copy.

Capitol Reader offered short executive summaries of political books to people with too much to read and too little time: exactly the sort of material likely to be useful to politicians, political staff, pundits, journalists, lobbyists, and policy professionals. Its later catalogue included a 2017 handbook for “new and seasoned politicos.”

So let us be precise. The politically consequential habit of delegating reading did not begin when someone opened a chat window. It existed when powerful people bought briefing books, executive summaries, and condensed versions of arguments they did not have time, or perhaps inclination, to read in full.

However, in all fairness, I read the crib notes to my own book, Outfoxed. I wasn’t impressed, though at least they had the decency to make their errors in print.

The real issue

There is, of course, a real criticism to make.

An AI summary can be wrong. It can omit the crucial qualification, invent a detail, flatten an author’s style, erase uncertainty, and confidently tell you that chapter seven contains something that never happened. It is particularly dangerous when readers cannot distinguish between a text-grounded explanation and an answer manufactured from the model’s general statistical fog.

That is not an argument against AI-assisted reading. It is an argument for using it competently.

Ask for quotations. Ask it to identify the chapter and page supporting a claim. Compare its answer with the actual text. Use it to explain, question, contextualize, challenge, and locate, not to issue a certificate declaring that you have “read” a book you have not read.

In other words, treat it the way a responsible person should have treated every CliffsNotes booklet, editorial, executive briefing, political summary, lecture handout, and book review ever printed.

The joke Futurism missed

The delicious irony is that the publication criticizing AI for mediating a reader’s relationship with a book exists in a media environment built on mediation. Journalism itself routinely compresses books, studies, court documents, reports, hearings, speeches, and technical material for an audience that will not encounter the source in full.

That function can be valuable. It is often essential.

But it is strange to say that a journalist writing a 900-word account of a 300-page report is performing a public service, while a reader asking an AI to clarify a confusing chapter is surrendering their brain to the machines.

The difference is not that one involves a summary and the other does not. The difference is that one is an older, professionally respectable intermediary and the other is a new, unfashionable, highly visible one.

AI did not invent intellectual shortcuts. It made the shortcut answer back.

Leave a Reply

Your email address will not be published. Required fields are marked *