Your dreams are not random: what a 2026 AI study actually found
You were at work, except the corridor turned into a hospital ward and your manager had a school friend's face. It feels random. A study published in Communications Psychology on April 28, 2026 suggests otherwise: dreams rework waking life through who you are and what you are living through.
Quick answer
Researchers used AI language models to score more than 3,000 dream and waking reports. Dreams did not replay the day: they turned it into more visual, spatial, social and bizarre scenes. Mind-wandering, how much people value dreams, sleep quality and the 2020 lockdown were all linked to what people dreamed. It describes group trends; it does not decode any single dream.
What did the researchers analyze?
A team led by Valentina Elce at the IMT School for Advanced Studies Lucca asked 207 adults to describe their dreams and their waking experiences every day for two weeks. The main dataset, collected between 2020 and 2024, holds 3,366 reports. A second set of 351 dream reports came from the spring 2020 lockdown.
Participants also completed tests on sleep, personality and cognition. Three large language models scored each report on 16 dimensions, from emotional intensity to bizarreness. According to the institute, they did about as well as human raters.
What patterns did they find?
Compared with waking reports, dreams moved away from thoughts about oneself toward scenes you perceive: places, movement, several characters, odd events. Work, school or hospital settings did show up, but blended into new scenes rather than copied.
Individual traits mattered. People prone to mind-wandering reported more fragmented, fast-changing dreams. Those who valued dreams reported richer, more immersive ones. Sleep quality was linked to dream content too. During lockdown, dreams carried more emotional intensity and more references to limits, effects that faded in the following years.
These are statistical links across a group of volunteers. “Not random” does not mean every image hides a message, and the study cannot tell you what your dream means.
What to note in your dream journal
The study suggests looking at the shape of a dream, not only its plot. When you wake up, ask: was it mostly images or mostly thoughts? Did the setting keep changing? Could you act, or were you blocked? How strong was the emotion? And how did you sleep?
Fictional example: “Office that became a train station. I could not find the exit. Anxious, 4/5. Slept badly, deadline Friday.” One entry proves nothing. Ten entries can show whether the blocked exit comes back on short nights or before deadlines. Our dream journal guide and our article on recurring dreams go further.
Why dream data deserves care
Dream reports can touch on worries, relationships or health. Before using an AI dream tool, check what it stores and whether you can delete it. See our privacy policy.
Noctalia does not hand you a single meaning. It helps you keep and reread your own dreams, so the connections stay yours.
Frequently asked questions
Are dreams random?
Not entirely. In this 2026 study, dream content was linked to personal traits and recent events. That does not mean every detail carries a hidden message.
Does AI prove what a dream means?
No. AI can detect patterns in language, themes and emotion, but meaning still depends on context and the dreamer's life.
What should I record after waking?
The scene, the emotion and its strength, people and places, setting changes, whether you could act, and how you slept.
Sources and further reading
- Elce et al. (2026), “Individual traits and experiences predict the content of dreams”, Communications Psychology, 4, 69
- IMT School for Advanced Studies Lucca, “What shapes the content of our dreams?”
Updated October 9, 2026
Read next
More resources on memory, dreams and AI
Last updated: