Functional Medicine

Does AI Contribute to Insomnia?

In his sonnet “To Sleep,” Keats pleads to be delivered from his “curious Conscience, that still lords / Its strength for darkness, burrowing like a mole.” The mole is provisioned by day: emails, social media, AI all day, every day. I have spent most of my career studying insomnia, though largely in other people. These past three weeks, the burrowing was mine. My restless mind probed what AI could do and cross-examined every answer, from business branding to whether my XERF beauty treatment would justify its cost. My sleep contracted from 7.5 hours to 6, and the deficit left my mind trudging through molasses. I weighed the usual suspects: perimenopause, travel, my husband away on business. Intuition insisted on another culprit. So I turned to the medical databases.

The short answer. Heavy use of generative AI is linked to poorer sleep. In a 2026 study of 868 university students, the worst sleepers scored highest on AI dependence, and night owls scored highest overall (Demirhan et al., 2026). The likely mechanism is well documented: demanding mental work in the evening raises cognitive arousal, which delays and disrupts sleep (Wuyts et al., 2012; Kalmbach et al., 2020). The antidote is timing and ritual. Give AI your mornings, set it a curfew at night, and signal to your nervous system that the day is over: a book or a few minutes of breathwork.

What the 2026 study on AI and sleep found

Published in the International Journal of Mental Health and Addiction, the study is among the first to examine generative AI use alongside sleep and the body clock (Demirhan et al., 2026).

The findings

  • Poorer sleep, greater dependence. The worse a student slept, the higher their AI dependence score.
  • Night owls scored highest. The link held after accounting for age, gender, and sleep quality.
  • Gender made no difference.

The wider research on screens points the same way. Among 122,058 adults, screen use in the hour before bed was linked to a 33% higher prevalence of poor sleep and roughly 50 minutes less sleep a week (Zhong et al., 2025), and a National Sleep Foundation panel reached consensus that screen use before sleep impairs sleep health in young people (Hartstein et al., 2024). AI is a newer screen, and frankly a more absorbing one.

Why AI work keeps the mind awake

Research on AI and sleep is in its infancy, but earlier sleep research is telling. Sleep scientists call the racing, problem-solving mind at bedtime cognitive arousal. In one study, 30 minutes of demanding tasks before bed delayed sleep and increased high-frequency brain activity- the rhythm of an alert, thinking brain- in early deep sleep, the stage where slow delta waves should prevail (Wuyts et al., 2012). Consider this: 30 minutes of problem-solving alters sleep architecture. What is your wearable telling you about your deep sleep on nights you use AI?

Why interactive use is worse

AI asks you to respond, decide, and respond again. In young people filmed with wearable cameras, every 10 minutes of interactive screen use in bed cost about 9 minutes of sleep (Brosnan et al., 2024). That means 30 minutes with AI in bed costs you almost 30 minutes of sleep. Does this sound familiar?

What sustained mental work does to the nervous system

Heart rate variability (HRV), a window onto the nervous system, falls with stress (Kim et al., 2018) and with sustained mental work: during a one-hour computer task, it dropped markedly, most steeply in the first 15 minutes (Delliaux et al., 2019).

The disappearing pause

Work used to have natural pauses: waiting on a reply, walking to a meeting, drafting by hand. Each let the nervous system descend from a small peak of effort. AI compresses them. An answer arrives in seconds. You can fit a day’s decisions into 30 minutes. In my view, what keeps us awake is the sheer volume of processing, which AI makes endless.

Night owls are the most vulnerable

Night owls scored highest for AI dependence, and their second wind, a late surge of alertness known as the wake maintenance zone, arrives later and stays longer: hours that AI fills effortlessly (Demirhan et al., 2026; de Zeeuw et al., 2018). If you are a night owl, protect your late evening as you would an important meeting, and save demanding AI work for the morning.

My evening protocol for heavy AI days

I use AI every day for my practice, my research, and my personal life. What I have changed is when I use it. I keep a strict cut-off: no AI after 9 p.m. I still answer emails, but I don’t use AI. At 10 p.m., I choose one thing that downregulates my nervous system: a book, some breathwork, or my PEMF mat. And no, I don’t do them all. I am asleep between 10:45 and 11 p.m.

My bedtime ritual

  1. Write tomorrow’s to-do list on paper.
  2. Switch the phone to airplane mode.
  3. Take Sleep Select by Moss Nutrition to help you get deeper sleep. Use code J064 at www.mossnutrition.com.
  4. Five minutes of slow belly breathing.

As for the XERF, I went ahead. It was worth it.

This article is for education and is not a substitute for individual medical advice. If your insomnia persists, please book an appointment with me.

Frequently asked questions

Does using AI affect sleep? Heavy AI use is linked to poorer sleep (Demirhan et al., 2026).

Can AI change brain waves? Indirectly. Demanding mental work before bed raises fast brain activity in deep sleep (Wuyts et al., 2012); direct research on AI is still to come.

How long before bed should I stop using AI? Two hours. Cutting evening screen time helps you fall asleep earlier and sleep longer (Perrault et al., 2019).

Is AI the problem or screens, in general? Both, but interactive use, AI included, depletes more sleep than passive viewing (Brosnan et al., 2024).

Does belly breathing help you fall asleep? Yes! Slow breathing before bed improves self-reported sleep (Eide et al., 2026) and shortens sleep onset in people with insomnia (Tsai et al., 2015).

What is the simplest change I can make tonight? Write tomorrow’s to-do list on paper (Scullin et al., 2018).

References

  1. Brosnan B, Haszard JJ, Meredith-Jones KA, et al. Screen use at bedtime and sleep duration and quality among youths. JAMA Pediatr. 2024. doi:10.1001/jamapediatrics.2024.2914
  2. de Zeeuw J, Wisniewski S, Papakonstantinou A, et al. The alerting effect of the wake maintenance zone during 40 hours of sleep deprivation. Sci Rep. 2018;8:11012. doi:10.1038/s41598-018-29380-z
  3. Delliaux S, Delaforge A, Deharo JC, Chaumet G. Mental workload alters heart rate variability, lowering non-linear dynamics. Front Physiol. 2019;10:565. doi:10.3389/fphys.2019.00565
  4. Demirhan E, et al. The role of biological rhythms in generative AI addiction: associations with chronotype and sleep quality. Int J Ment Health Addict. 2026. doi:10.1007/s11469-026-01679-3
  5. Eide E, et al. Slow breathing techniques before bedtime and the effects on sleep: a systematic review. Sleep Med Rev. 2026. doi:10.1016/j.smrv.2026.102284
  6. Hartstein LE, Mathew GM, Reichenberger DA, et al. The impact of screen use on sleep health across the lifespan: a National Sleep Foundation consensus statement. Sleep Health. 2024;10(4):373-384. doi:10.1016/j.sleh.2024.05.001
  7. Kalmbach DA, Buysse DJ, Cheng P, Roth T, Yang A, Drake CL. Nocturnal cognitive arousal is associated with objective sleep disturbance and indicators of physiologic hyperarousal in good sleepers and individuals with insomnia disorder. Sleep Med. 2020;71:151-160. doi:10.1016/j.sleep.2019.11.1184
  8. Kim HG, Cheon EJ, Bai DS, Lee YH, Koo BH. Stress and heart rate variability: a meta-analysis and review of the literature. Psychiatry Investig. 2018;15(3):235-245. doi:10.30773/pi.2017.08.17
  9. Perrault AA, Bayer L, Peuvrier M, et al. Reducing the use of screen electronic devices in the evening is associated with improved sleep and daytime vigilance in adolescents. Sleep. 2019;42(9):zsz125. doi:10.1093/sleep/zsz125
  10. Scullin MK, Krueger ML, Ballard HK, Pruett N, Bliwise DL. The effects of bedtime writing on difficulty falling asleep: a polysomnographic study comparing to-do lists and completed activity lists. J Exp Psychol Gen. 2018;147(1):139-146. doi:10.1037/xge0000374
  11. Tsai HJ, Kuo TBJ, Lee GS, Yang CCH. Efficacy of paced breathing for insomnia: enhances vagal activity and improves sleep quality. Psychophysiology. 2015;52(3):388-396. doi:10.1111/psyp.12333
  12. Wuyts J, De Valck E, Vandekerckhove M, et al. The influence of pre-sleep cognitive arousal on sleep onset processes. Int J Psychophysiol. 2012;83(1):8-15. doi:10.1016/j.ijpsycho.2011.09.016
  13. Zhong C, Masters M, Donzella SM, Diver WR, Patel AV. Electronic screen use and sleep duration and timing in adults. JAMA Netw Open. 2025;8(3):e252493. doi:10.1001/jamanetworkopen.2025.2493