This is a bonus post, following from yesterday’s, in which I interviewed the artificial intelligence known as ChatGPT about meditation.
This time I gave the AI the prompt, “Write a Buddhist sutta, in the style of the Pali canon, in which the Buddha tells Ananda to stop playing Candy Crush on his iPhone.”
Here’s what ChatGPT came up with.
Once, the Blessed One was dwelling at Sāvatthī in Jeta’s Grove, Anāthapiṇḍika’s Park. And at that time, the Venerable Ananda was addicted to playing Candy Crush on his iPhone.
The Blessed One approached Ananda and said, “Ananda, what are you doing?”
Ananda replied, “I am playing Candy Crush, Lord.”
The Buddha spoke: “Ananda, this is not the path to liberation. The playing of such games only leads to craving and attachment. To achieve true peace and wisdom, one must let go of such distractions and engage in virtuous and meaningful actions.”
Ananda listened to the Buddha’s words and put down his iPhone, vowing to let go of his attachment to Candy Crush. He turned his attention to the Buddha’s teachings and soon realized the path to enlightenment.
And so it is said:
“Just as a bird is freed from its cage
when the door is opened, so too one is freed
from the bonds of ignorance and craving
by practicing the path taught by the Buddha.”
Appropriately, the image above was generated by another AI (Craiyon: formerly DALL-E Mini), using the prompt, “book illustration of the Buddha holding an iPhone.”
Coming back down to earth, ChatGPT’s predecessor was prone to using vile racist and misogynistic language, because its database of information was drawn from the web, and the web is full of that sort of stuff. For the latest version of the AI, a program had to be developed that could recognize and screen out the toxic material. In order to do that, Kenyan workers were paid as little as $1.32 an hour (via a subcontractor of OpenAI, which owns ChatGPT) to read and categorize the text. Workers were traumatized by reading, for example, graphic descriptions of bestiality and pedophile sex.
The contract between OpenAI and the subcontractor ended after an article was published showing that Facebook had employed the subcontractor to screen content. In this case, Kenyans were required to view images and videos of executions, rape and child abuse for as little as $1.50 per hour. This is traumatizing work.
AI and social media are fun, but there’s a dark and exploitative side to it as well.
Wow. This is the kind of reporting I wish happened in news outlets. I know this post isn’t reporting, but it’s the real news in a balanced way. It’s also very interesting and warm hearted. Thank you for being so on fire! (Is there even mindful and balanced reporting anywhere? Maybe sorta Louis Theroux?)
I highly recommend reading this very original and lucid essay by Ted Chiang on what it is that ChatGPT is actually doing, and its very real limitations.