The Connection Between Nightmares, Health and AI Art

The word nightmare originally referred to an evil spirit thought to afflict its victims while they slept. The 13th-century Norse book the Ynglinga Saga describes a king of Upsal, Vanlande, being killed by just such an evil spirit, stating, “But when he had slept but a little while he cried out, saying that the Mara was treading upon him.”

Who has nightmares?

Nightmares are a common part of the human experience. Between 50% and 85% of adults occasionally have them. Adult females report more nightmares than their male counterparts. Children 3-6  have nightmares more frequently than adults but nightmares are common among people of all ages. 

Humans are not the only animals who dream. Researchers have been studying animal dreaming since at least 2001. Pet-owners have long speculated that animals experience nightmares. “He’ll start twitching a lot while dreaming, suddenly wake up, and finds me wherever I am and cuddles hard!,” writes Kiki_lou on Reddit about their cat. In a 2015 study, scientists even observed evidence of nightmares in rats.

Nightmares are common among humans and animals, but what causes nightmares to form?

Why do we have nightmares?

Mayo Clinic lists stress or anxiety, trauma, certain drugs, mental health disorders, sleep deprivation or the consumption of media with frightening content as possible causes of nightmare disorder. According to the CDC, one in three adults do not get the recommended seven hours of sleep per night. College and high school students in particular are often accused of being sleep-deprived. 

In 2002, researchers at St. Joseph’s University in Philadelphia linked nightmares in elementary-aged children to feelings of anxiety. The International COVID-19 Sleep Study indicates that the pandemic had a major effect on dreaming. Nightmares are also symptoms of depression and PTSD. 

AI and nightmares

Kelly Bulkeley, Director of the Sleep and Dream Database, suggested in a Psychology Today article that AI could play a role in dream interpretation in the future. Bulkeley asked three of his friends to describe a vivid image from their dreams and used those descriptions as prompts for a computer to generate several images. The results received mixed responses with one woman being particularly disturbed by the images that came from her prompt.

Images generated by artificial intelligence can often be described as dreamlike and even nightmare-ish. AI generator, Deapdream seems to embrace the comparison.  Amelia Winger-Bearskin, of the Digital Worlds Institute, University of Florida explains that AI image generators don’t think in the same way that people do. These generators use an algorithm to compare the pixel values of thousands of images. According to Winger-Bearskin, they don’t actually know that dogs “usually have two eyes, and typically only one face per head.”

Now it’s time to look to the future and see how artificial intelligence can impact the way we interpret nightmares.