Internet and computer effects

Sparrow et al (Reference Sparrow, Liu and Wegner2011) reported one of the first studies to point toward an impairing effect of Internet use on memory, and their results have been prominently cited both in the scientific literature (over 1400 citations according to Google Scholar, June 2021) and in some of the popular press articles that raise alarm bells regarding the negative impact of the Internet on memory. Sparrow et al reported that when given a difficult general information question (eg, What country has a national flag that is not rectangular?), people automatically think of computers and the Internet. Their evidence for this conclusion came from a clever modification of the well-known Stroop task, in which people are slower to name the colour of a printed word when the colour is incongruent with the printed word (eg, the word ‘Green’ printed in blue) than when it is congruent (eg, the word ‘Blue’ printed in blue). Sparrow et al reported that participants were slower to name the colour of computer-related words (eg, Google, Internet) than non-computer-related words (eg, Nike, Yoplait). They suggested that this interference with colour naming indicates that people automatically think of computers and the Internet when given difficult general knowledge questions rather than searching their memories. Moreover, they found that participants showed poorer memory for trivia items that they typed into a computer when they were told that the computer would save this information than when they were instructed that the computer would delete it. However, participants did frequently remember where they saved the answers (ie, which folder) even when they failed to recall the information itself.

These findings raised the possibility that automatic reliance on computers and the Internet is having negative effects on memory. However, Marsh and Rajaram (Reference Marsh and Rajaram2019) noted that they failed to find the same kind of memory-impairing effects observed by Sparrow et al (Reference Sparrow, Liu and Wegner2011), and Kahn and Martinez (Reference Kahn and Martinez2020) also failed to observe such effects. Camerer et al (Reference Camerer2018) reported a failure to replicate the so-called Google Stroop effect, ie, the interference effect in the modified Stroop task. Sparrow (Reference Sparrow2018) pointed out several differences between the replication and the original study, but a subsequent replication study by Hesselman (Reference Hesselman2020) that attempted to address those differences also failed to find evidence for the Google Stroop effect.

Moreover, Storm and Stone (Reference Storm and Stone2015) reported three experiments showing a beneficial effect of relying on a computer to store information: when participants saved a pdf that contained a word list to the computer rather than attempting to memorise the words themselves, they later recalled more items from a different list of words in a second pdf presented shortly after the first pdf. This saving-enhanced memory was not observed when participants were instructed that the saving process ‘was potentially fallible and that sometimes they might save a file on the computer only to later find it unavailable for restudy (2015, 185)’. The saving-enhanced benefit also failed to emerge when the saved information did not constitute a significant memory burden (eg, a two-word list; for related research, see Giebl et al Reference Giebl, Mena, Storm, Bjork and Bjork2020). Runge et al (Reference Runge, Frings and Tempel2019, Reference Runge, Frings and Tempel2020) replicated the saving-enhanced effect on memory for word lists, showed that it can be observed for motor materials (ie, sequences of finger movements; Runge et al Reference Runge, Frings and Tempel2020), and demonstrated that benefits of saving on subsequent performance extend to cognitive tasks such as solving arithmetic problems (Runge et al Reference Runge, Frings and Tempel2019). However, Runge et al (Reference Runge, Frings and Tempel2020) also documented a mnemonic cost for the saved material: participants recalled fewer of the items that they were able to save to a computer than those that they could not save. There are other conditions in which offloading items to external storage can impair memory for the offloaded items on a delayed memory test, but people who are aware that there will be a delayed test can largely overcome the detrimental effects of offloading on long-term memory (Grinschgl et al Reference Grinschgl, Papenmeier and Meyerhoff2021).

Taken together, these studies reveal a mixed pattern of results concerning retention of information over time that includes computer-related impairments, enhancements, and null effects. What seems clear, however, is that the extant data do not support far-reaching claims that the Internet or computers are ‘killing’ or ‘ruining’ memory.

GPS effects

The impact of technology on memory has also been assessed in studies concerned with the effects of GPS and similar navigational aids. Behavioural studies have shown that reliance on a navigational aid impairs memory for the route in a virtual navigation task (eg, Gardony et al Reference Gardony, Brunyé and Taylor2015). Additional evidence comes from an fMRI study by Javadi et al (Reference Javadi2017). Prior to scanning, participants toured the Soho area of London and learned about its spatial layout. During scanning the following day, participants were shown a filmed simulation of Soho and engaged in two types of spatial navigation tasks: one in which they had to rely on their memories from the previous day to plan how they would navigate, and one in which they followed navigation instructions provided by the experimenters, similar to navigating based on GPS. In the condition that required memory-based navigation, there was increased activity in the hippocampus, which numerous previous studies have linked to memory and spatial navigation, and also in a part of the prefrontal cortex that had been implicated in spatial planning by earlier work. By contrast, these increases were not observed when participants relied on experimenter-provided navigation instructions that functioned as a GPS.

Because some previous research has linked hippocampal volume with the active use of spatial navigation and memory abilities (Konishi and Bohbot Reference Konishi and Bohbot2013; Maguire et al Reference Maguire2000; Woollett and Maguire Reference Woollett and Maguire2011; but see also, Weisberg et al Reference Weisberg, Newcombe and Chatterjee2019), the results reported by Javadi et al (Reference Javadi2017) are consistent with behavioural findings that relying on GPS negatively impacts memory of a navigated route, and raise the possibility that it would negatively impact spatial memory more generally. More direct evidence on this point comes from a study by Dahmani and Bohbot (Reference Dahmani and Bohbot2020), who examined spatial memory and navigation performance as a function of lifetime GPS experience, as assessed by a questionnaire, in a sample of 50 drivers. On two different virtual maze tasks that required spatial learning without GPS, they found that individuals with greater reported lifetime GPS experience showed impaired spatial navigation and memory compared to individuals with less lifetime GPS experience. While such a finding could indicate an impairing effect of GPS use on spatial memory, it could also be that individuals who rely more on GPS do so because they have poor spatial navigation and memory skills. However, Dahmani and Bohbot failed to find any relationship between GPS use and self-reported sense of direction, and they also reported effects of GPS use on spatial memory even after accounting for the effects of subjective sense of direction. In addition, Dahmani and Bohbot performed a follow-up assessment three years after initial testing with a subset of 13 participants, and found that more reliance on GPS since initial testing was associated with a steeper decline in spatial memory at follow-up. Thus, the results of Dahmani and Bohbot provide some empirical support for the idea that reliance on GPS can impair spatial memory.

Photo-related effects

A third area in which issues related to the effects of technology on transience have emerged concerns the impact of photographs on memory. Given that online platforms such as Facebook and Instagram are overrun with photos, questions concerning how taking and reviewing photographs impact memory are timely and receiving increasing attention. Research in my lab conducted during the 1990s revealed that when participants reviewed photographs of specific incidents that occurred in an earlier videotaped sequence of actions, subsequent memory for those events was improved relative to a no-review condition (Koutstaal et al Reference Koutstaal, Schacter, Johnson, Angell and Gross1998). More recently, benefits of photo review on subsequent memory have been documented in studies in which healthy and brain-injured participants are equipped with wearable cameras such as SenseCam, which allow people to record and later review everyday experiences (eg, Khachatoorian et al Reference Khachatoorian, Loveday, Dima, Illingworth, Conway and Hanschel2021; for reviews, see Harvey et al Reference Harvey, Langheinrich and Ward2016; Silva et al Reference Silva, Pinho, Macedo and Moulin2018). However, the benefits of photo review can be accompanied by a cost. For example, in a study by Koutstaal et al (Reference Koutstaal, Schacter, Johnson and Galluccio1999), after carrying out simple actions in the lab (eg, pounding a nail into a block of wood), participants reviewed photos of some of the actions, which boosted recall of those actions on a later test. But recall of non-reviewed actions on that same test was reduced compared to a condition in which participants did not review any photographs, suggesting that selective review of photographs can impair memory for non-reviewed content (see Cinel et al Reference Cinel, Mack and Ward2018; Henkel and Milliken Reference Henkel and Milliken2020 for recent evidence illustrating the benefits and costs of reviewing and editing photos on memory).

Researchers have also begun to investigate how the act of taking a photo impacts memory for the photographed event. Henkel (Reference Henkel2014) reported a study in which participants toured a museum and were instructed to take photos of some objects and simply look at others. When they had taken a photo of an object during the tour, participants remembered fewer objects and their locations on a later memory test than when they had just looked at an object. Similar results were reported by Tamir et al (Reference Tamir, Templeton, Ward and Zaki2018) in a study where participants toured a church; some participants had cameras and were instructed to take as many photos as they wanted, whereas others toured the church without a camera and took photos. Participants who took photos recalled fewer details of the tour a week later compared with participants who did not.

One possibility is that these impairing effects of taking photos on later memory are attributable to cognitive offloading (Risko and Gilbert Reference Risko and Gilbert2016), where the participants rely on the photo as a kind of external memory and thus do minimal memory-related processing. Evidence against this idea was obtained in a study by Soares and Storm (Reference Soares and Storm2018) in which participants went on a virtual museum tour: they found an impairing effect of taking photos on later memory in a condition where participants deleted their photos shortly after taking them. These findings raise the possibility that, rather than cognitive offloading, the activities in which people engage when taking a photo, such as finding the best angle or lighting conditions for taking the photo, could reduce encoding of the object or event and thus impair subsequent memory. However, other evidence indicates that impaired memory is not a necessary consequence of taking a photo of an event: Henkel (Reference Henkel2014) reported that the photo-impairment effect was not observed when participants were instructed to zoom in on a specific part of the photographed object. In addition, Barasch et al (Reference Barasch, Diehl, Silverman and Zauberman2017) reported a photo-enhancement effect on subsequent memory for visual aspects of a museum tour when participants were able to decide for themselves when to take a photo. Taken together, the emerging evidence does not support the simple conclusion that taking a photo necessarily impairs memory for the photographed object or event, but instead paints a more complex picture, indicating that effects on memory depend on such factors as an individual's goals when taking a photo and what aspects of memory for an event are tested (for more detailed review and discussion, see Henkel et al Reference Henkel, Nash, Paton, Lane and Atchley2021).

Summary and conceptual distinctions

To summarise, then, there is some evidence from each of the three areas reviewed (Internet, GPS, and photos) that these forms of technology can impair the ability to retain memories over time, but questions remain concerning the reliability of some of the evidence (ie, Internet and computer effects), and there are also data from research on computer effects and photos that highlight conditions under which aspects of memory may be enhanced.

At a more general conceptual level, however, I believe that discussions regarding the effects of the Internet, GPS, and photos on memory are muddled by a failure to specify exactly what is meant when claims such as ‘Internet use is killing your memory’ are made. Specifically, I suggest that it is important to distinguish among at least three different ways in which technology or media might impair memory (or potentially enhance it; but for the present purposes, the focus is on impairment). I will refer to these as task-specific, domain-specific, and domain-general effects (for discussion of similar conceptual issues regarding near- and far-transfer in brain training programmes, see Simons et al Reference Simons2016). A task-specific effect would occur when relying on technology impairs memory for the specific task performed, but does not have broader effects. Thus, for example, a task-specific GPS effect would occur if relying on GPS resulted in poor memory for the specific route on which one used GPS, but had no effects on one's memory beyond that specific route. By contrast, a domain-specific effect would occur when relying on technology impairs memory for both the specific task performed and other tasks within a particular domain, but not on tasks outside that domain (by ‘domain’, I am referring broadly to classes of information and tasks, such as spatial, verbal, or visual, as well as to broadly accepted distinctions among forms of memory, such as working memory, episodic memory, and semantic memory). Continuing with the GPS example, a domain-specific effect would occur if relying on GPS impaired memory for a specific route in which one relied on GPS, and also impaired the ability to establish memories for other routes and spatial layouts even when not using GPS, but did not impact memory abilities for non-spatial information. Finally, a domain-general effect would occur if relying on technology impairs memory broadly for tasks other than the one relied on and in multiple domains of memory. In this scenario, sufficiently extensive reliance on GPS, for instance, would compromise one's ability to retain both spatial information and non-spatial information, such as verbal/linguistic information or non-spatial aspects of everyday events (eg, people and objects associated with a specific event).

Viewing the previously discussed studies from the perspective of the aforementioned distinctions, I think that there is solid evidence for an impairing effect of technology at a task-specific level, such as reduced memory for a photographed item (Henkel Reference Henkel2014; Tamir et al Reference Tamir, Templeton, Ward and Zaki2018) and impaired memory for routes on which one used GPS (Dahmani and Bohbot Reference Dahmani and Bohbot2020; Gardony et al Reference Gardony, Brunyé and Taylor2015). However, I would characterise the evidence for a domain-specific effect as only suggestive and restricted to the studies of GPS, ie, poor spatial memory tentatively linked to lifetime GPS use in one sample of study participants (Dahmani and Bohbot Reference Dahmani and Bohbot2020). Finally, I see no evidence for a domain-general effect of technology on memory in any of the studies discussed here. Yet that is the kind of broad impairment typically implied by worries that technology may be ‘killing’ or ‘ruining’ memory. Going forward, it will be interesting to see whether such evidence emerges.

Mind wandering and media multitasking

Cognitive studies have long established that dividing attention during encoding, typically by requiring participants to perform an attention-demanding task in addition to encoding target information (eg, a list of words), severely impairs an individual's ability to form a new memory (eg, Craik et al Reference Craik, Govoni, Naveh-Benjamin and Anderson1996; and see Gardony et al Reference Gardony, Brunyé and Taylor2015, for evidence that divided attention contributes to impaired memory for a route following GPS use). A growing number of studies have also demonstrated negative effects on memory as a consequence of spontaneous shifts of attention during encoding that occur as a result of mind wandering – ie, when people focus on thoughts that are unrelated to the primary task at hand (for reviews of mind-wandering research, see Smallwood and Schooler Reference Smallwood and Schooler2006; Seli et al Reference Seli, Risko, Smilek and Schacter2016). Mind wandering is particularly relevant to educational settings, as a number of studies have demonstrated that it occurs frequently during both classroom and online lectures, and that engaging in mind wandering can impair retention of lecture content (Lindquist and McLean Reference Lindquist and McLean2011; Risko et al Reference Risko, Anderson, Sarwal, Engelhardt and Kingstone2012; Szpunar et al Reference Szpunar, Kahn and Schacter2013; for review, see Schacter and Szpunar Reference Schacter and Szpunar2015). Of particular relevance to this article, negative effects on memory have also been observed as a consequence of media multitasking, ie, diverting attention during a lecture by using a smartphone or laptop. Wammes et al (Reference Wammes, Ralph, Mills, Bosch, Duncan and Smilek2019) reported two studies showing that in a live university lecture setting: (a) students often engaged in media multitasking, (b) the frequency of media multitasking increased during the course of the lecture, and (c) more frequent media multitasking was associated with worse retention on quizzes and tests.

Linking back to the distinctions I suggested in the discussion of transience, the findings of Wammes et al (Reference Wammes, Ralph, Mills, Bosch, Duncan and Smilek2019) reflect a task-specific effect of media multitasking on subsequent memory for content from a particular lecture, and indeed, other earlier studies had shown similar task-specific effects of media multitasking on memory in educational contexts (eg, Rosen et al Reference Rosen, Carrier and Cheever2013; Wood et al Reference Wood, Zivcakova, Gentile, Archer, De Pasquale and Nosko2012). However, we can also ask whether there are additional domain-specific or domain-general effects of media multitasking. For example, do individuals who regularly engage in media multitasking develop chronic deficits such that they exhibit memory impairments even when they are not engaging in media multitasking on a particular task? Do such deficits show up in only one domain (eg, educational materials) or are they evident across multiple domains of memory function?

A growing number of studies have examined the relationship between chronic media multitasking and memory performance, and the general finding is that higher levels of media multitasking are often, but not always, associated with poorer memory performance on a variety of tasks – even under task conditions in which participants do not engage in media multitasking (eg, Uncapher et al Reference Uncapher, Thieu and Wagner2016; for review, see Uncapher and Wagner Reference Uncapher and Wagner2018). As discussed by Uncapher and Wagner (Reference Uncapher and Wagner2018), these effects are most likely attributable to lapses of attention that reduce memory performance. A recent study by Madore et al (Reference Madore2020) provides strong evidence on this point. They reported that attentional lapses during the retrieval phase of a laboratory memory task, revealed by measures of pupil diameter and brain electrical activity that have been reliably associated with momentary lapses of attention, were associated with poorer memory performance on the retrieval task. Moreover, responses on a media multitasking questionnaire revealed that higher levels of self-reported media multitasking were associated with increased attentional lapses, leading to worse memory performance during the retrieval task.

The findings of Madore et al (Reference Madore2020) and those of previous studies that have linked high levels of media multitasking with poor memory performance raise the possibility that chronic reliance on media multitasking causes impaired memory. Given that these memory deficits have been observed during tasks on which people are not engaging in media multitasking, the data further suggest a possible domain-general memory impairment. However, as Madore et al (Reference Madore2020), Uncapher et al (Reference Uncapher, Thieu and Wagner2016), and Uncapher and Wagner (Reference Uncapher and Wagner2018) are careful to point out, the extant data are correlational, and thus do not allow us to determine definitively whether media multitasking causes attentional failures that result in memory impairments, or whether individuals who start out with poor attention and memory abilities are especially prone to engaging in media multitasking. Sorting out these two possibilities is an important issue to be addressed by future research. It will also be important to assess whether different kinds of media multitasking are more or less likely to lead to absent-minded memory failures (Baumgartner and Wiradhany Reference Baumgartner and Wiradhany2021).

Prospective memory and cognitive offloading

While concerns that media multitasking might increase attentional failures that produce absent-minded memory errors are justified, it is also important to note that absent-minded errors can be reduced substantially by reliance on external memory aids (ie, cognitive offloading), and that modern media and technology offer effective ways of delivering such aids. For example, research on prospective memory – remembering to carry out intentions in the future – has shown that absent-minded memory errors often occur when cues to carry out an intended action are not present at the time a future action needs to be executed (eg, McDaniel and Einstein Reference McDaniel and Einstein2007). Creating and relying on external reminders significantly improved performance in a naturalistic prospective memory task (Gilbert Reference Gilbert2015). A recent meta-analysis of prospective memory studies in middle-aged and older adults revealed overall positive outcomes for external memory aids in enhancing prospective memory (Jones et al Reference Jones, Benge and Scullin2021). Commonly used external memory aids include smartphones, smartwatches, and Google Calendar, which have proven useful for improving prospective memory performance in individuals with brain injuries and disease who are especially prone to absent-minded forgetting (eg, El Haj et al Reference El Haj, Moustafa, Gallouj and Allain2020; Jamieson et al Reference Jamieson2019; Svoboda et al Reference Svoboda, Richards, Leach and Mertens2012).

A particularly affecting example of a catastrophic absent-minded memory failure that can be remedied by external memory aids, including some made possible by modern technology, concerns a phenomenon that occurs regularly every summer: well-intentioned parents forget that they have left their infant in a hot car, resulting in the death of the infant. These tragic cases typically result from an unfortunate confluence of circumstances that conspire to produce a seemingly impossible type of absent-minded forgetting, including absorption with personal concerns, reliance on automatic behaviour, and – most critically – lack of a retrieval cue at the moment it is needed to remind a parent that their infant is in the back seat (for more extensive discussion, see Reference SchacterSchacter in press, chapt 2; Weingarten Reference Weingarten2009). However, effective external reminder systems using modern technology have been developed that can deliver the missing retrieval cue. For example, a device called the Elepho eClip can be attached to a child's car seat and sends both visual and auditory alerts to smartphones in order to remind drivers that there is a baby in the car (see Baldwin Reference Baldwin2019).

The main takeaway from this brief overview of external reminders is that technology can be used to help overcome the annoying and sometimes dangerous consequences of absent-mindedness. Thus, while there are legitimate concerns regarding the negative impact of activities such as media multitasking on memory failures that result from attentional lapses, these concerns should be balanced by the realisation that smartphones and related devices that contribute to media multitasking can also play a positive role in combating absent-minded forgetting.