How content licensing solves AI's source quality crisis

Around 1589, an English clergyman named William Lee asked Queen Elizabeth I for a patent. He’d invented a machine that could knit stockings, and he was eager to have the royal seal of approval. But when she saw Lee’s device, the Monarch is said to have replied “thou aimest high, Master Lee. Consider thou what this invention could do to my poor subjects. It would assuredly bring to them ruin by depriving them of employment, thus making them beggars.”

While social media can foster connectivity and creativity, it can also be a powerful vector for the spread of false information. In a 2005 survey by the Pew Research Group, only 5% of adults reported using at least one social media platform. By 2021, that number had grown to 72%.

The ancient Romans had a secret weapon. Hidden away in the Temple of Jupiter on the Capitoline Hill was a collection of oracular pronouncements written in Greek hexametric verse known as the Sibylline Books. In a world that seemed to be governed by unseen and capricious forces, they offered much-needed certainty. The Romans took comfort in the fact that, when things seemed to be going wrong, the Books would show them what they had to do in order to make things better, whether that was establishing a festival in honor of the goddess Flora or burying Greeks and Gauls alive under the forum. For modern content creators, it can feel like we’re in the same boat as the Romans.

YouTube recently wanted me to watch a video about the ancient Egyptian bureaucrat-turned-demigod Imhotep. I knew it was going to be a hot mess from the thumbnail alone. Obviously AI-generated, it made poor Imhotep look like he’d just visited the face-painting booth at the county fair. The actual content, such as it was, featured an artificial voice reading out nonsensical statements like “Imhotep lived in the ancient Egyptian period 267 BCE AD 648” (he actually lived around the 27th century BCE). Unfortunately, schlock like this has become all too common as many creators try to churn out as much content as possible with the least amount of effort. But while the Imhotep video was just cringy, there is a darker side to low-effort content. YouTuber Harry Brewis (aka hbomberguy) recently put out a four-hour video exploring the increasing prevalence of plagiarized content on the platform.

The other day I stumbled upon an interview that POLITICO did with Fr. Paolo Benanti, a Franciscan friar who advises Pope Francis on matters related to artificial intelligence. Until I read that article, I had no idea that the Vatican was paying such close attention to AI. But I’m glad they do because they can offer a unique and valuable perspective that’s often overlooked. Benanti’s core argument is that we need to have “human-centric” AI that operates within carefully prescribed limits. As he told POLITICO, “[s]ome people treat AIs like idols, like oracles, like demigods. The risk is that they delegate critical thinking and decisional power to these machines.”