advertisement
advertisement

How this optics research scientist is finding real humor in artificial intelligence

How this optics research scientist is finding real humor in artificial intelligence
[Illustration: Artur Tenczyński]


Janelle Shane is an optics research scientist by day and a humorist by night, documenting her experiments with machine learning on a blog called AI Weirdness. In a typical project, she’ll train a neural network (a type of algorithm inspired by human brains) on a particular kind of data (craft beer names, say, or knitting patterns) then let the AI generate suggestions. These range from plausible to nonsensical, but the funniest fall in between: new names for paint colors included shades like Stoner Blue and Turdly (a light mustard). Shane’s hobby blossomed into editorial collaborations last year with CNN and The New York Times, for which she trained an AI to come up with new Halloween costumes (Zombie Schoolgirl, Donald McDonald), and she has a book due out November 5. The title derives from an experiment in which she taught an AI to write pickup lines: You Look Like a Thing and I Love You: How AI Works and Why It’s Making the World a Weirder Place. The book, she says, explores what algorithms “are good at, what they’re not good at, and how we should think about their place in our world.”

advertisement
advertisement