The extended mind: are smartphones becoming part of human memory?
How external memory, smartphones, search engines and AI assistants influence remembering, thinking and decision-making: what the research shows, what it doesn't, and how to use your digital memory well.
Try to recall the phone number of the last person who called you. Now think about how many numbers, birthdays, routes, passwords and photos your phone holds for you, and how rarely you have to remember them yourself. In 2008 the philosopher David Chalmers wrote that the iPhone he had bought a month earlier had "already taken over some of the central functions of my brain".1 The question in the title is meant literally. If a device stores our memories, retrieves our facts and increasingly does some of our thinking, is it becoming part of the mind? And what happens to the memory that is still in our heads? This guide works through the evidence: the philosophy of the extended mind, what the famous "Google effect" really showed, photos, GPS and reminders, the phone-on-the-desk debate, AI assistants and the fear of "digital dementia". It ends with practical habits for using your digital memory well.
01Cognitive offloading: an old habit on a new device
Psychologists have a name for storing information outside your head: cognitive offloading. Evan Risko and Sam Gilbert define it as using physical action to change the information-processing demands of a task so as to reduce cognitive demand. Their examples range from tilting your head to read a rotated image to programming a smartphone to remind you of an appointment.2 Shopping lists, sticky notes and saved files all count.
What is new is not offloading itself but its scale. A single device now carries your contacts, calendar, photos, notes, maps and search history; it is almost always within reach; and, with AI assistants, it can summarise, draft and recommend as well as store. That turns an old philosophical puzzle into an everyday one.
02The extended mind: when a notebook counts as memory
In 1998 the philosophers Andy Clark and David Chalmers published a short paper called "The Extended Mind", built around a thought experiment. Inga hears about an exhibition at the Museum of Modern Art, recalls that the museum is on 53rd Street and walks there. Otto, who has Alzheimer's disease, carries a notebook everywhere; he hears the same news, looks up the address in his notebook and walks there too. Clark and Chalmers argued that the notebook plays the same role for Otto that biological memory plays for Inga, and proposed what they called the parity principle:3
"If, as we confront some task, a part of the world functions as a process which, were it done in the head, we would have no hesitation in recognizing as part of the cognitive process, then that part of the world is (so we claim) part of the cognitive process."
They did not claim that any tool counts. Otto's notebook qualifies because of four features: it is a constant in his life, its information is directly available without difficulty, he automatically endorses what he retrieves from it, and he consciously endorsed the information when he wrote it down.3 Run a smartphone through the same test and the answer is mixed:
- Constant and available: yes. For many people the phone is closer to hand than Otto's notebook ever was.
- Automatically endorsed: often, and this is where it gets interesting. We tend to trust what the screen shows us, including search results and AI answers we never checked.
- Endorsed in the past: only partly. Your own notes and photos pass; a search result or a chatbot answer written by someone, or something, else does not.
That is why many philosophers now take a middle position. Richard Heersmink and John Sutton argue that most web applications are better understood as scaffolds for memory and thinking than as fully integrated parts of the mind, although some personalised applications on wearable devices may already be deeply integrated.4
Psychology offers a second lens: transactive memory, Daniel Wegner's term for the way people in close relationships share out the work of remembering. In a 1991 study of 118 people in dating relationships, real couples remembered more than pairs made up of people from different couples when left to their own division of labour, but less when the researchers imposed a division on them, which suggests the couples were relying on an established system of who remembers what.5 On this view your phone is less a piece of your brain than a new member of your memory team.
03What the "Google effect" really shows
The most famous study in this area, often summed up as the "Google effect" or "digital amnesia", appeared in Science in 2011. Betsy Sparrow, Jenny Liu and Daniel Wegner reported four studies suggesting that people faced with difficult questions are primed to think about computers, and that when people expect to have future access to information, they remember less of the information itself and more about where to find it. They concluded that the internet had become a primary form of external, or transactive, memory.6
The paper became shorthand for "Google is making us forget", but its record is more mixed than the headlines. Its first finding, a reaction-time test suggesting that hard questions bring computer-related words to mind, was one of 21 social-science experiments from Nature and Science re-run by a large replication project in 2018, and it did not replicate;7 a separate replication in 2020 also found no conclusive evidence for it.8 Across that 2018 project as a whole, 13 of the 21 findings replicated, with effects averaging about half their original size.7
The memory side of the story fits better with later work from other labs:
- Saving frees space. In three experiments, saving one file before studying a new one improved memory for the new file, but not when the saving process was unreliable.9 Offloading can be a strategic way to clear your head, provided you trust the store.
- Searching becomes a habit. People who used Google to answer difficult trivia questions were later more likely to reach for Google even for a new set of easy questions.10
- Access feels like knowledge. Across nine experiments, searching the internet for explanations made people rate their own knowledge higher, as if what they had looked up were already in their heads.11
04Photos, maps and reminders: offloading in daily life
The same trade shows up in ordinary habits, and the details matter.
Photos
In a 2014 museum study, Linda Henkel found that people who photographed objects remembered fewer of them, and fewer details about them, than people who only looked; zooming in to photograph a specific detail removed the effect.12 A 2017 series of studies added a twist: people who were free to take photos during an experience recognised more of what they saw but less of what they heard.13 The camera is less a memory thief than an attention director.
Navigation
Finding your way is the clearest case of a skill that may weaken when it is outsourced, in this case to satellite navigation. Louisa Dahmani and Véronique Bohbot assessed 50 regular drivers and found that those with more lifetime GPS experience had worse spatial memory when they had to navigate on their own. Thirteen of them were retested three years later, and greater GPS use over that period was associated with a steeper decline in hippocampal-dependent spatial memory; the heavier users had not turned to GPS because they felt they had a poor sense of direction.14 The follow-up sample was small and the study observational, so treat it as a warning sign rather than proof.
Reminders
Not all offloading is a loss. Using external reminders for future intentions, such as taking medication or sending a form, is highly effective. Research on this "intention offloading" also shows that people's choices about when to set reminders are guided by how much they trust their own memory, with systematic biases that stay stable over time.15 The skill worth building is knowing when to offload, not refusing to do it.
05The phone on the desk: attention, not just memory
In 2017 Adrian Ward and colleagues tested whether the mere presence of your own phone drains your mind. In the first of two experiments, 520 smartphone users left their phones on the desk, in a pocket or bag, or in another room. Available working memory capacity and fluid intelligence fell as the phone got closer and more visible, even though participants did not report thinking about their phones; a second experiment, with 275 participants, found the cost was largest for the people most dependent on their phones.16
The "brain drain" idea spread quickly, but it has not held up cleanly. A pre-registered direct replication of the second experiment, published in 2022, found no difference between phone locations on either task.17 A 2023 meta-analysis of 22 studies found a small overall negative effect of smartphone use and presence (g = −0.14) that was significant only for memory, varied by region, and was close to zero in North American studies.18
What is better established is the cost of interruptions. In a 2015 experiment, phone notifications alone, without participants touching their phones, significantly disrupted performance on an attention-demanding task, by an amount comparable to actually using a phone for calls or texts.19
The evidence that a silent phone in your bag drains your mind is weak. The evidence that a buzzing one costs you attention is much stronger.
06AI assistants: from storing facts to outsourcing thinking
Search engines changed where facts live. AI assistants go further: they summarise, draft, explain and recommend, so what gets offloaded is no longer just memory but parts of reasoning and judgement. The research is young, but three recent studies point the same way.
- Learning. In a field experiment with nearly a thousand high-school maths students, access to GPT-4 while solving practice problems raised grades by 48% with a standard ChatGPT-style interface and by 127% with a tutor designed to safeguard learning. When access was taken away, students who had used the standard version did 17% worse than students who never had access, while the safeguarded tutor largely avoided that harm. Without guardrails, the authors write, students used the AI as a "crutch".20
- Writing and memory. In an MIT Media Lab study released as a preprint, 54 participants wrote essays using ChatGPT, using a search engine or using no tools while their brain activity was recorded with EEG. Brain connectivity was strongest without tools, moderate with search and weakest with the chatbot; chatbot users also reported the lowest sense of ownership of their essays and struggled to accurately quote their own work.21 It is a small study of one task, so treat it as an early signal rather than a verdict.
- Work. A survey of 319 knowledge workers by researchers at Microsoft Research and Carnegie Mellon, covering 936 real examples of using generative AI at work, found that higher confidence in the AI was associated with less critical thinking, while higher confidence in one's own abilities was associated with more. Critical thinking did not disappear; it shifted towards verifying information, integrating responses and overseeing the task.22 The findings are self-reported.
For decisions, the older research on automation is sobering. A review of studies on automated decision aids found that when the aid is imperfect, people make both omission errors (missing problems the aid failed to flag) and commission errors (following its wrong advice); this automation bias affects experts as well as novices, and training or instructions do not prevent it.23
Here the extended mind thesis carries a sharp warning. One of the conditions that makes Otto's notebook part of his mind is that he automatically trusts it.3 Automatic trust is reasonable for a notebook you wrote yourself. It is a risk when the "notebook" writes itself, can be confidently wrong, and sounds equally sure either way.
07Is "digital dementia" real?
The phrase "digital dementia" suggests that a lifetime of screens wears down the brain. The best test so far points the other way. A 2025 meta-analysis in Nature Human Behaviour pooled 57 studies covering 411,430 adults over 50, with an average age of 68.7. Using digital technology was associated with a lower risk of cognitive impairment (odds ratio 0.42) and slower cognitive decline over time (hazard ratio 0.74), and the associations held after accounting for demographic, socioeconomic and health factors and for proxies of cognitive reserve.24
The authors call this "technological reserve" and are careful about causation: the studies are observational, and more work is needed to test whether the relationship runs in both directions, for example because people with healthier brains use more technology.24 The fair summary is that there is no good evidence that smartphones cause dementia, while there is evidence that specific skills you stop practising, such as finding your own way, can fade.14
08How to use your extended mind well
The research points to a few habits that keep the benefits of offloading without paying the hidden costs:
- Decide what to offload and what to own. Let the phone hold what you only need to look up: numbers, dates, documents. For knowledge you need to think with, at work or in study, practise retrieving it. In a classic study, testing yourself produced substantially better retention two days and a week later than rereading did, even though rereading felt more effective.25
- Treat looked-up answers as borrowed, not owned. Searching can make you feel you know more than you do,11 so before a decision that matters, check the source and ask whether you could explain it without the screen.
- Use AI as a tutor, not a crutch. Try the problem first, ask for hints rather than answers, and ask the assistant to quiz you; a tutor designed this way protected learning in the maths experiment.20
- Verify before you act. Imperfect automated advice leads to both missed problems and followed mistakes, even among experts.23 Make checking a habit for anything consequential, especially AI summaries of documents you have not read.
- Photograph with intention. Take the picture, then keep looking, or zoom in on the detail you care about; focused photography did not impair memory in the museum study.12
- Navigate on your own sometimes. Look at the whole route before you set off and do familiar trips without turn-by-turn directions, so your spatial memory stays in use.14
- Cut the interruptions. Turn off non-essential notifications and batch the rest; notifications alone can cost attention.19 For deep work, put the phone in another room: the evidence for a "brain drain" is mixed, but the habit costs nothing.1718
- Keep a backup brain. Know a few critical phone numbers by heart and keep key documents accessible offline. An extended mind can be lost, locked or out of battery.
09Where OCXLY lands
Are smartphones becoming part of human memory? In a practical sense, yes. They hold what we used to carry in our heads, and our memory has adapted by keeping track of where things are rather than what they say. That is not obviously a decline: people have always shared remembering with partners, colleagues, books and notebooks, and offloading done well frees the mind for other work.59
The risks lie elsewhere: in trust that is automatic rather than earned, in the illusion that information we can reach is knowledge we have, in interruptions that fragment attention and, with AI, in handing over the thinking along with the remembering. The safeguarded maths tutor shows that design choices can change the outcome.20 Our view: judge a tool by whether it leaves you more capable when the screen is off, not just more productive while it is on. For more on the mental cost of constant connection, see our deep dive on digital cognitive fatigue, and for the emotional side of the feed, social media and self-esteem.
10Frequently asked questions
Is my smartphone part of my mind?
It depends on the theory. Under Clark and Chalmers' extended mind thesis, a tool counts when it is constantly used, easy to access, automatically trusted and filled with information you endorsed, which fits your own notes better than search results.3 Many researchers prefer to call web tools scaffolds for thinking rather than parts of the mind.4
Does Google make you forget things?
Not exactly. A 2011 study suggested that when people expect to find information later, they remember where it is better than what it says. Part of that paper later failed to replicate, but related studies show that saving and searching change what people choose to remember.6710
Do smartphones cause "digital dementia"?
There is no good evidence that they do. A 2025 meta-analysis of more than 411,000 older adults found technology use associated with a lower risk of cognitive impairment, although observational studies cannot prove cause and effect.24
Does using ChatGPT make you worse at thinking?
Early evidence suggests it depends on how you use it. Unrestricted AI help raised students' practice scores but lowered their later performance without it, while a tutor-style design avoided most of that harm.20 Studies of essay writers and knowledge workers link leaning on AI with lower engagement and, among people who trust it most, less critical thinking.2122
Does GPS harm your sense of direction?
Heavy GPS use has been linked to worse spatial memory when navigating without it and, in a small follow-up, to a steeper decline over three years.14 Studying the full route and navigating familiar trips yourself keeps the skill in use.
How can I use my phone without weakening my memory?
Offload storage but practise retrieving what matters, turn off non-essential notifications, verify important answers, and use AI to quiz you rather than to answer for you.251920