Back to news
AnalysisJune 5, 2026· 3 min read

Your attention span dropped to 47 seconds. AI bots are making it worse.

Psychologist Gloria Mark tracked focus time over 20 years: 2.5 minutes in 2003, 75 seconds by 2012, 47 seconds by 2020. Now AI is eroding the cognitive effort we need to think clearly.

Our Take

Mark's attention-span data is real and concerning, but her AI cognitive-atrophy argument rests on a reasonable hypothesis, not yet on the kind of long-term evidence she herself is waiting for.

Why it matters

If depth of processing—the effort required to understand and retain information—is genuinely declining as we offload thinking to AI, the downstream effects on critical thinking and decision-making could be substantial. Schools and enterprises need to know whether this is already happening or still speculative.

Do this week

Engineering leaders: audit which tasks your teams have delegated to Claude or ChatGPT in the past six months, then require one peer review or independent verification per task to rebuild the cognitive friction that prevents atrophy.

Attention spans have collapsed over two decades

Gloria Mark, a psychologist at UC Irvine who has studied digital technology use for 30 years, tracked focus time in what she calls "living laboratories." Using sensors and heart-rate monitors, she measured how long people could concentrate on a single task before switching attention.

In 2003, the average was 2.5 minutes. By 2012, it had dropped to 75 seconds. Between 2014 and 2020, it fell again to 47 seconds on average. The decline correlates with measurable stress: Mark observed direct correlation between rapid attention-switching and elevated heart rates in her study subjects.

Mark also found that frequent task-switching extends completion time and harms emotional well-being. The research aligns with findings from lawsuits against Meta and Google, in which school districts have accused social media platforms of designing addictive products that harm student mental health. Around 1,200 school districts are pursuing similar legal action.

The picture is not entirely one-directional. A 2024 survey of LGBTQ+ teenagers found that while some experienced social media as harmful, others reported it as a place of belonging and identity formation. Mark acknowledges that long-term evidence on social media's net effects on children remains inconclusive despite decades of research.

AI may accelerate cognitive decline through outsourcing

Mark expressed particular concern about AI chatbots like ChatGPT, Claude, and Gemini. When we ask these tools to write, summarize, or evaluate content, we avoid what she calls "depth of processing"—the active engagement with information that leads to learning and retention.

"You're deferring your cognitive work to AI," Mark said. "And it's not good for us." Her worry is that critical thinking capacity will weaken over time, similar to how muscles atrophy without use. People with diminished critical thinking are more vulnerable to misinformation.

Synthetic companion AI poses a different but related risk. Human relationships require effort, time, and emotional understanding. Interaction with a sycophantic bot requires none of these, potentially eroding emotional intelligence—a skill already declining in some cohorts by self-reported measures.

Mark summarized the compound risk: "If we continue on this trajectory, attention spans are diminished, loneliness is rising, boredom is rising, emotional intelligence decreasing, and our sense of purpose is also decreasing."

Rebuild deliberate cognitive effort into daily routines

Mark advocates for what she calls "creating new life routines" that prioritize effort and friction. Reading a full article rather than relying on a summary, meeting friends in person instead of over text, navigating without GPS when feasible—these are not anti-technology moves, she argues, but deliberate choices to exercise cognitive faculties.

"The more effort we put into something, the deeper the satisfaction we stand to gain," Mark said. She is not arguing for abandoning AI or digital tools—"we can't give it up." Rather, she emphasizes intentionality about when and how we use them.

This is particularly relevant for knowledge workers and teams relying heavily on AI for drafting, analysis, and decision support. The question is not whether to use these tools, but which cognitive steps remain worth doing by hand to preserve judgment and learning.

#AI Ethics#LLM#Research
Share:
Keep reading

Related stories