What the data actually shows
The foundational study is Sparrow, Liu, and Wegner (2011), which described what's popularly called the "Google effect" or "digital amnesia." Across several experiments, people who expected to be able to look information up later remembered the information less well but remembered where to find it better. The brain appeared to treat the computer as an external store and to prioritise the path to the information over the information itself.
This connects to an older, well-established idea: transactive memory. Daniel Wegner's work described how people in close groups — couples, teams — naturally divide up who remembers what, so the group as a whole knows more than any individual, with each person also tracking who to ask. The internet, on this view, becomes a transactive memory partner: we offload to it and remember how to retrieve, much as we'd remember that a partner keeps track of the finances.
Cognitive offloading more broadly — using calculators, notes, search — is well documented and often beneficial: it frees limited working memory for other tasks and lets people accomplish far more than unaided memory would allow. But studies also suggest costs. Offloaded information is less likely to be retained internally, and some research finds that habitual reliance on external sources can reduce the effort people put into remembering or reasoning themselves. The evidence is still developing and contested in places, so strong claims in either direction outrun the data.
Why this feels different from how it actually is
It feels like the internet is either making us geniuses or rotting our brains because both narratives are vivid and neither captures a quieter middle. The gains in access are obvious and immediate; the costs to retention are slow, invisible, and only noticed when you reach for a fact you once knew and find it's gone. The two operate on different timescales, which makes balanced judgment hard.
It also feels different because we conflate "knowing" with "being able to find out." When information is always a search away, the felt boundary between what's in your head and what's on your screen blurs — people routinely overestimate how much they actually know, mistaking easy access for personal understanding. The web can inflate your sense of your own knowledge precisely because it makes looking things up so frictionless.
And the loss of retention rarely announces itself, because the system mostly works. You don't notice the facts you've offloaded until the connection drops or the source disappears. So the tradeoff stays hidden under normal conditions, which is part of why the debate swings between extremes rather than settling on "it's a reallocation with costs and benefits."
That's not stupidity; it's a reallocation — but it's a reallocation with tradeoffs.
What the research says to do about it
The research on learning suggests that what you want to retain, you have to engage with effortfully — looking something up is not the same as learning it. Practices like retrieval practice (testing yourself rather than re-reading) and spacing repetition over time are among the most robustly supported ways to move information from "findable" to "known." If retention matters for a given thing, the internet is a starting point, not a substitute for that effort.
Being deliberate about what you offload and what you internalise is the practical core. Offloading routine, look-up-able facts frees capacity, which is genuinely useful; internalising the concepts and frameworks you reason with day to day is what lets you think without a search bar. The skill is choosing which is which, rather than treating everything as either worth memorising or safe to outsource.
Treating search as a transactive partner — knowing what it's good for and what it can't do — is itself a form of competence. Strong understanding still depends on having enough internal knowledge to ask good questions, judge sources, and connect ideas, none of which the internet supplies on its own.
What the research says does not help
Panicking that the internet is destroying memory or intelligence outruns the evidence. The studies describe a shift in what we remember and some costs to retention, not a documented decline in general cognitive ability — and offloading to external tools has been part of human cognition since writing. Catastrophising treats a tradeoff as a collapse.
The opposite error — assuming that because everything is findable, internalising knowledge no longer matters — also fails. Easy access inflates how much people think they understand, and you cannot reason well, spot errors, or ask good questions about a topic you carry none of in your own head. Access is not the same as understanding.
Relying on having looked something up as if that were learning it doesn't build retention. The Google-effect work suggests that expecting to find information later actively reduces how well you remember it, so passive look-up, repeated, is close to the opposite of the effortful engagement that retention requires.
You cannot reason well, spot errors, or ask good questions about a topic you carry none of in your own head.
What this looks like in real life
Remembering the path, not the fact
In the Sparrow, Liu, and Wegner experiments, people who expected to be able to look information up later remembered the information itself less well but remembered where to find it better. The brain treated the computer as an external store and prioritised the route to the facts over the facts themselves — a reallocation, not stupidity.
Feeling like you know more than you do
When a topic is always a search away, the boundary between what's in your head and what's on your screen blurs, and people routinely overestimate their own knowledge. The web can inflate your sense of understanding precisely because looking things up is so frictionless — until the connection drops or the source disappears and you reach for a fact that isn't there.
Real numbers in context
There isn't a clean number for "how much smarter or dumber" the internet makes us, and you should distrust anyone who supplies one — this is a question about the structure of memory, not a single measurable IQ shift. What the experimental work does suggest is a directional pattern: when people expect information to stay accessible, they remember it less and remember its location more (Sparrow et al., 2011). That's a reallocation of what's retained, not a verdict on intelligence.
The useful framing is access versus retention. On access, the gain is enormous and uncontroversial — vastly more information reachable in seconds than any prior generation had. On retention, the evidence points to more offloading and some cost to what we carry internally, alongside the well-established benefit that offloading frees limited working memory. The net effect depends on which capacity you're measuring, which is exactly why "smarter or dumber" is the wrong question.