Users have noticed that the remoteli.io twitter chatbot, usually faithful to its cheerful messaging promoting remote work, can be subverted with a carefully worded user prompt.

Users were able to get the chatbot to claim responsibility for terrorist attacks, threaten the President, meow at other twitter users, print snippets

For a couple of years now I've been using neural networks to generate daily drawing prompts. With today's text-generating neural networks far too large to finetune on a list of existing prompts, I've turned to other methods. One method that works surprisingly well is

Humans are weird. The question is, do AI training datasets capture the full scope of our weirdness?

I decided to see if GPT-3 could successfully predict other human quirks when prompted with a short list of them. Here's what I gave GPT-3 DaVinci to complete:

Human things that