The Attention Evidence Gap (Inside my Advanced Topics in AI Law and Policy Class #2.2)
When your experience says one thing and the research says "it depends"
8:30 AM, Monday. Welcome back to Professor Farahany’s Advanced Topics in AI Law and Policy Class. Hopefully, you caught up on Week 1 of class. But if you haven’t, start with week 1, which are classes 1, 1.2, and 1.3.
We are now into Week 2, and this is class 2.2. If you haven’t taken class 2 yet, do that first. This semester, we break up the classes into three parts, Mondays, Wednesday, and Fridays. And I’m hoping you’ve finished your attention audit before joining us today.
As returning students know, we are building on our Fall semester AI Law and Policy class. At the very least, make sure you’ve covered AI Fundamental (What is AI), and How AI Actually works before embarking on the materials for this semester.
This week, we’re examining the capture of your attention by digital technologies. An auspicious time to be doing so given the launch of the historic trial against social media platforms. Jacob Ward has a terrific write-up of it that you should check out.
What the Science Actually Shows
On Monday, I shared what happened when students spent three days tracking their attention—disabling algorithmic feeds on Day 2, then enabling everything on Day 3.
The experiences were vivid. “Muscle memory.” “I didn’t feel in control at all.” “I would open Instagram and lose an hour.” One student dreaded Day 3 more than Day 2.
But the question we are exploring today, is Does the science support what they (and you) felt?
Hold onto your answer, and let’s see if the research changes your mind.
You’ve probably heard some version of this story:
Screens are destroying our attention spans
Kids can’t focus anymore
Social media is rewiring our brains
We’re becoming addicted
This narrative is everywhere. From bestsellers like The Shallows, to news articles, and worried parents. And after what students experienced in the Attention Audit, it feels true.
But what does the research actually show?
Lodge & Harrison: “It’s Complicated”
I assigned students a 2019 review article from the Yale Journal of Biology and Medicine, The Role of Attention in Learning in the Digital Age.
The authors, Jason Lodge and William Harrison (from the School of Education in The University of Queensland, Australia), examine the claim that digital technology is negatively impacting attention and learning.
Their starting point is the popular concern:
“Many commentators argue that access to the Internet is having a persistent detrimental impact on the brain. In particular, attention has been implicated as a cognitive function that has been negatively impacted by use of digital technologies for learning.”
Their conclusion?
“Across the two bodies of literature, a complex situation emerges placing doubt on the claim that the use of digital technologies for learning is negatively affecting the brain.”
That’s not “technology is fine.” It’s “it’s complicated.” (A law professor’s favorite mantra).
The neuroscience evidence is weak: Lodge and Harrison cite a review by Loh and Kanai: “although there are some examples of neuroscience studies pointing to changes in the brain as a result of Internet use, the evidence is far from conclusive.” No clear evidence of permanent “brain rewiring.”
Multitasking does hurt learning, but that’s not new: There’s “robust evidence of the negative impact of multitasking in digital environments on academic performance.” But multitasking has always hurt performance. This isn’t unique to digital technology. (I used to pass notes in class. But is that really the same as sending an instant message … constantly … throughout class?)
The reading/comprehension question is complicated: Some studies show worse comprehension on screens vs. paper. But factors like content type, study time, and demographics muddy the picture. “It remains uncertain as to when and how technologies influence reading and comprehension.”
The second assigned article, A Review of Evidence on the Role of Digital Technology in Shaping Attention and Cognitive Control in Children, by Maria Vedechkina and Francesca Borgonovi in Frontiers in Psychology, goes deeper into research on children. The central finding:
“The lack of scientific consensus on whether digital technologies are good or bad for children reflects that effects depend on users’ characteristics, the form digital technologies take, the circumstances in which use occurs and the interaction between the three factors.”
In other words: it depends (a law professor’s favorite answer to student questions).
User characteristics matter: Age, pre-existing conditions, socioeconomic background, individual differences in cognitive capacity all moderate the effects.
The form of technology matters: Television vs. video games vs. social media. Fast-paced content vs. educational content. Passive viewing vs. interactive engagement. The authors note that “action video games, in particular, seem to be associated with benefits in visual and spatial selective attention.”
Context matters: How it’s used (educational vs. entertainment), when (age of exposure, time of day), with whom (co-viewing with parents vs. alone), and what it displaces (high-quality interaction vs. nothing).
And on the “brain rewiring” claim specifically:
“Major brain changes or brain ‘rewiring’ as a product of screen exposure, social media, or internet use is considered to be highly unlikely.”
Why Is the Evidence So Mixed?
This is frustrating. You rightly want to know, is this stuff bad for us or not? Why can’t science give us a straight answer?
A few reasons:
Causation vs. correlation: Do screens cause attention problems, or do people with attention problems use more screens? Most studies can’t tell us.
Self-report problems: Heavy media multitaskers report more attention difficulties—but they don’t always show deficits on objective tests. As one review notes, heavy multitaskers may *choose* to engage differently rather than being *unable* to focus.
Technology changes faster than we can study it: The research literature “largely reflects outdated devices and patterns of use.” Studies from 2015 may not tell us much about TikTok in 2025.
Publication bias: Dramatic findings get published. Null results often don’t.
Heterogeneous effects: Averages may hide that some people are harmed while others benefit. “Some features of digital media may be particularly problematic, but only for certain users and only in certain contexts.”
If you did the Attention Audit, think about your own experience. Did you choose to engage differently on Day 3, or were you unable to disengage? That distinction—preference vs. impairment—is exactly what the science struggles to untangle.
If the science is mixed, how did we get here? Tim Wu’s, The Attention Merchants, provides historical context.
The chapter I assigned traces Google’s ideological evolution. In 1998, when Sergey Brin and Larry Page were graduate students, they wrote a paper arguing that advertising-funded search engines were inherently compromised. Ads would corrupt search results. The incentive to serve advertisers would undermine the incentive to serve users.
Their original vision for Google was a pure information utility. Search should give you the best results, period.
But Google needed revenue. The dot-com crash made this urgent. The solution was AdWords—but notice how Google framed it.
Google convinced itself that targeted advertising wasn’t corrupting; it was helpful. If you search for “running shoes,” an ad for running shoes isn’t an intrusion, it’s useful information. Traditional advertising interrupts what you’re doing. Google’s advertising serves what you’re doing.
This reframing allowed Google to maintain its self-image as a company that helps users while building an advertising empire.
Wu’s chapter traces the origins of the digital advertising model, how Google made it ideologically acceptable. But the chapter I assigned doesn’t get into what happened next.
Once advertising became the business model, the next step was inevitable: optimize engagement.
If your revenue depends on showing ads, you need eyeballs. If you need eyeballs, you need attention. If you need attention, you need to capture it.
The algorithmic feed that you and the in-person students experienced on Day 3? That’s the optimization layer. Google convinced itself that serving ads was helping users. The platforms that followed optimized for keeping you there so they could serve more ads.
The equation that emerged:
User Attention × Ad Impressions × Click-Through Rate × Conversion = Revenue
Every variable creates an incentive to capture more attention, more often, more intensely.
So here’s where we are:
You all felt something during the Attention Audit. Day 2 was different from Day 3.
The science can’t definitively tell us what that difference means for cognition or development.
The business model incentivizes attention capture regardless of what the science shows.
Google’s framing, that targeted advertising helps users, is still the dominant industry narrative. Platforms argue that algorithmic feeds give you what you want. That the personalization is a service.
But based on the Attention Audit experience: Is that true? Is the algorithmic feed giving you what you want? Or is it giving you what captures your attention, which may not be the same thing?
On Friday, we’ll ask, can the law help? And we’ll discover that the laws that exist don’t address the problem students actually experienced.
If the science is uncertain, and your experience suggests something is happening, what should law do?
Class Dismissed.
The entire class lecture is above, but if you’d like to support my work or go deeper in your learning, please upgrade to being a “paid subscriber.”
Paid subscribers also get access to class readings packs, discussion questions, bonus content, full archives, virtual chat-based office hours, additional readings, as well as one live Zoom-based class session per semester.
