Around 300 B.C., King Ptolemy I — the new ruler of Egypt and a former general of Alexander the Great — tasked an adviser with a modest mission: “to collect, if possible, all the books in the world.” Over the next two centuries, the great library in the Ptolemaic capital of Alexandria would be filled with hundreds of thousands of papyrus scrolls: the full corpus of ancient Greek and Egyptian literature along with Buddhist, Jewish, and Zoroastrian texts. Ships would be searched for books when they docked at Alexandria, and royal agents would pay hefty sums for almost any written work. A booming market in fakes and forgeries soon emerged. Entrepreneurial scribes dashed off scrolls of supposed secret wisdom from famous thinkers — one was titled Everything Thucydides Left Unsaid — while others created books that mixed the authentic with the imagined. In Alexandria’s merchant quarter, stalls that once sold vegetables and baskets were “replaced with those stacking rolls and rolls of books,” writes historian Islam Issa.
Eventually, the library had to hire experts to wade through the sea of bogus texts and identify genuine treasures. The web, our modern-day library of Alexandria, faces a similar problem. This digital repository of human knowledge is being swamped with AI-generated slop — pointless listicles, nonsensical how-to guides, and factually flawed news summaries churned out by content factories that want to grab clicks and ad revenue on the cheap. To save users the hassle of scrolling through reams of garbage links in its search engine, Google has now started showing users AI-generated answers to their queries. But those answers are sometimes wrong — one user who wanted a fix for a car’s faulty turn signal was advised to “replace the blinker fluid” — and pull traffic and dollars away from useful, human-run websites. Maybe the tech giant should hire more humans to curate trustworthy collections of knowledge. It could call them “librarians.”
This is the editor’s letter in the current issue of The Week magazine.