Coastal and Rural Digital Exclusion: The Case for Voice AI

Rory Baxter

1 a

, Oksana Hagen

2 b

, Amir Aly

2 c

, Ray B. Jones

1 d

and Katharine Willis

3 e

School of Nursing and Midwifery, University of Plymouth, Plymouth, U.K.

School of Engineering, Computing and Mathematics, University of Plymouth, Plymouth, U.K.

School of Art, Design and Architecture, University of Plymouth, Plymouth, U.K.

{rory.baxter, oksana.hagen, amir.aly, ray.jones, katharine.willis}@plymouth.ac.uk

Keywords:

Digital Exclusion, Digital Divide, Voice-Based Systems, Chatbots, Voice AI.

Abstract:

The adoption of digital platforms for health and social resources disadvantages vulnerable populations, in-

cluding older adults and those living in remote or deprived areas. This could be remediated using voice-based

conversational AI (Voice AI) systems delivered via landline phone, bypassing the requirement for digital de-

vice access or digital skills. British rural and coastal regions often have poorer digital infrastructure and

pockets of deprivation, and consequently higher levels of digital exclusion. This study explores digital exclu-

sion in Southwest England and the suitability of Voice AI systems for supporting digital inclusion. Seventeen

participants aged 50 years or over were interviewed by telephone and took part in one of two face to face

focus groups to identify how digital exclusion impacts access to health and wellbeing resources. The results

indicated that digital access was severely impacted by unreliable infrastructure and exacerbated by limited

digital skills. Phone-based Voice AI systems could then provide viable solutions to support access to digital

health and social resources for digitally marginalised coastal and rural communities.

1 INTRODUCTION

Digital exclusion comprises individuals being unable

to participate in society due to barriers to their dig-

ital access, such as limited digital skills, poor de-

vice access, or negative attitudes toward technology

(Helsper, 2021; van Dijk, 2020). It is often associ-

ated with a negative impact on health and wellbeing

through limiting access and adoption of digital health

and wellbeing resources (Mee et al., 2024). Older

adults are particularly vulnerable to digital exclusion

(Lythreatis et al., 2022) due to infrequent use of digi-

tal resources, old devices, and poor digital skills (Gal-

listl et al., 2021; Ueno et al., 2023). Despite these

barriers to digital access for older adults, health tech-

nologies can support wellbeing through, for exam-

ple, facilitating social connections (Quan-Haase et al.,

2017) and reducing social isolation and loneliness

(Hajek and K

onig, 2021). This means digitally ex-

cluded older adults face health inequalities by not

https://orcid.org/0000-0001-6057-6595

https://orcid.org/0000-0002-5486-7609

https://orcid.org/0000-0001-5169-0679

https://orcid.org/0000-0002-2963-3421

https://orcid.org/0000-0001-9988-1933

having consistent access to these resources. Commu-

nity services and groups are increasingly transitioning

to online platforms (Spanakis et al., 2021), leading to

increased social isolation for digitally excluded older

adults and consequently leading to poorer long-term

health outcomes (Goldman et al., 2023).

In the UK and Europe, a large factor in digital ex-

clusion is rurality (Salemink et al., 2017), exacerbat-

ing the difﬁculties faced by older adults who form a

higher proportion of the population in these regions

(Department for Environment Food and Rural Affairs,

2024). Within the UK, this typically results from

poorer infrastructure compared to urban areas (Philip

et al., 2017). Indeed, rural areas in the UK are more

likely to suffer difﬁculties in accessing a reliable in-

ternet connection compared to urban areas (Rural Ser-

vices Network, 2022). An additional complication is

that 3% of the UK rural landmass has no 4G signal

available and 9% of rural homes lack access to 4G

signal (Ge et al., 2022), limiting the utility of mobile

devices (Mascheroni and

Olafsson, 2016).

In the UK, coastal regions are associated with

higher levels of deprivation compared to other areas

(Mee et al., 2024). Furthermore, coastal areas may

be considered the ’last mile’ of infrastructure (Hen-

derson and Roche, 2020), meaning that there may be

734

Baxter, R., Hagen, O., Aly, A., Jones, R. B. and Willis, K.

Coastal and Rural Digital Exclusion: The Case for Voice AI.

DOI: 10.5220/0013259300003911

Paper published under CC license (CC BY-NC-ND 4.0)

In Proceedings of the 18th International Joint Conference on Biomedical Engineering Systems and Technologies (BIOSTEC 2025) - Volume 2: HEALTHINF, pages 734-741

ISBN: 978-989-758-731-3; ISSN: 2184-4305

logistical issues that affect the development of model

infrastructure. The intersectionality of coastal and ru-

ral regions often compounds these issues; for exam-

ple, Cornwall, in the Southwest of the UK, has the

highest average age and higher levels of deprivation

compared to other areas of the country (Ministry of

Housing Communities and Local Government, 2019).

1.1 Models of Digital Exclusion

Contemporary models of digital exclusion (Helsper,

2021; van Dijk, 2020) present a hierarchical approach

to understanding the factors that lead to digital exclu-

sion. Van Dijk’s (2020) model in particular presents a

comprehensive overview of how the levels of the dig-

ital divide impact an individual’s wider participation

in society. Core components of the model are the four

factors that comprise digital access, which in turn in-

ﬂuence the outcomes of digital engagement.

The ﬁrst level is Motivation and attitude which

is driven by needs, motives, and attitudes. These in-

ﬂuence decisions to purchase digital hardware and en-

gage with digital resources. The absence of motiva-

tion to engage with technology can consequently lead

to unsatisfactory outcomes from digital media.

The second level is Physical access, which distin-

guishes between three types of access: physical, ma-

terial, and conditional. Physical access refers to an in-

dividual’s opportunities to use digital media privately

or in community settings such as libraries or commu-

nity centres. Material access comprises the resources

required to support continuous access to digital me-

dia, including subscriptions, software, and utilities

such as electricity. Finally, conditional access refers

to provisional access to applications or resources that

might be associated with digital media.

Th third level comprises Digital skills, i.e. com-

petencies, and literacy required to use digital media.

This encompasses both direct interaction with appli-

cations as well as with other people through digital

communications. The fourth level of access is Usage

inequality which comprises the difference in usage

between users. Some individuals may use digital me-

dia extensively, whereas others may restrict their us-

age due to their access to digital devices or skills. This

is typically explored across frequency and amount of

use, use diversity, and activity of use.

In Van Dijk’s (2020) model these four levels inﬂu-

ence an individual’s Outcomes of digital media use.

Engagement with digital media can provide beneﬁts

across several domains, including social, cultural, and

personal domains (Helsper et al., 2015). The four lev-

els of digital access inﬂuence the extent to which dig-

ital engagement is positive or negative.

Van Dijk’s (2020) model also includes factors that

represent an individual’s resources, and categorical

inequalities that have been shown to inﬂuence digi-

tal exclusion, which include personal (e.g. age) and

positional (e.g. region) factors. Personal categorical

inequality factors have a greater inﬂuence on digital

usage than positional factors, in part due to positional

factors also being inﬂuenced by the outcomes of par-

ticipation in a digital society (van Dijk, 2020).

1.2 Voice AI System as a Digital

Inclusion Tool

To tackle digital exclusion in coastal and rural re-

gions, solutions are needed that can overcome barri-

ers to digital inclusion. The ICONIC research project

(Intergenerational Codesign Of Novel technologies

In Coastal communities; see Jones et al., 2024) is co-

designing a voice-based conversational AI system

- Voice AI - delivered via a phone line. Co-design

embeds users within the design process, centering the

design of new technologies around their life experi-

ences to ensure that their needs are understood and not

assumed (Thabrew et al., 2018). Co-designed tech-

nologies are often perceived as more accessible, are

more likely to be adopted (Mitchell et al., 2023) and

can support digital inclusion by lowering the digital

skill barrier. The Voice AI would provide access to

online information or services in response to voice

commands via a phone line, bypassing the need for

devices, an internet connection, and advanced digital

skills.

Using voice interfaces to improve digital access is

not new, as it provides an alternative to traditional vi-

sual interfaces (Pradhan et al., 2018). Recent work

shows that conversational systems could be powerful

tools for social inclusion (Song, 2023), and tackling

digital exclusion (Sin et al., 2022). These systems

provide utility for older adults when searching for

information online (Pradhan et al., 2020), however,

when delivered via traditional platforms such as smart

speakers, smartphones, or computer, digital skills and

device access remain as barriers (Park and Humphry,

2019). As the ICONIC project’s device agnostic sys-

tem bypasses these equipment barriers, it might then

be an accessible solution to people living in coastal

and rural settings. A co-design approach will be sen-

sitive to the perceptions and attitudes of the users, en-

suring that user voices are centred within the design

process, supporting user acceptability and long-term

engagement (Thabrew et al., 2018).

Coastal and Rural Digital Exclusion: The Case for Voice AI

735

1.3 Research Questions

To understand the potential utility of phone-based

Voice AI systems in coastal and rural communities,

it is important to identify the factors that cause digital

exclusion. This study had two research questions. In

coastal and rural communities in Southwest England:

• What factors lead to digital exclusion in a rural,

coastal community?

• Could Voice AI systems support digital inclusion?

2 METHODOLOGY

2.1 Design

This study is part of the ICONIC research project

(Jones et al., 2024) which is taking an intergenera-

tional approach to address digital exclusion faced by

older people and the digital economic/employment

exclusion of younger people. The project recruited

participants in coastal and rural communities in

Southwest England, to join workshops focusing on

the co-design of a Voice AI system delivered via a

phone line. Participants took part in a telephone

semi-structured interview and a focus group discus-

sion held during a co-design workshop. The inter-

view was focused on understanding the participant’s

digital exclusion, and the focus group explored barri-

ers to accessing online information and services, and

opinions on Voice AI technology. Phenomenological

inquiry was selected as the research methodology, as

it provides an understanding of digital exclusion from

the perspective of participants, based on their lived

experience (Money et al., 2024). Ethics approval was

obtained from the University of Plymouth Arts and

Humanities Research Ethics and Integrity Committee

(project ID 3941).

2.2 Participants

The study took place with residents of a rural village,

Pendeen, and a coastal town, St Austell. Pendeen was

selected due to its isolation, situated in the far South-

west of England. It has 920 residents and is classiﬁed

as a ”Rural hamlet and isolated dwellings in a sparse

setting” and it is one of the most deprived places in

the UK (Ministry of Housing Communities and Lo-

cal Government, 2019). St Austell is also located in

the Southwest of England. It has 20,900 residents and

is one of the 10% most deprived locations in the UK

(Ministry of Housing Communities and Local Gov-

ernment, 2019). Eleven participants (10 female, 1

male) with a mean age of 71.5 (SD = 5.6) were re-

cruited via the Pendeen community centre. A further

six participants (all female) with a mean age of 61.2

(SD = 8.2) were recruited in St Austell, via social me-

dia, ICONIC recruitment partners (including [Rural

Town] Healthcare), and word of mouth. All seven-

teen participants were interviewed by telephone and

took part in one of two focus groups.

2.3 Data Collection

Interviews and focus groups were facilitated by mem-

bers of the research team. Semi-structured interviews

were held over the phone, as participants preferred

this platform to other options such as Zoom. For one

participant, the telephone interview had to be resched-

uled twice due to poor connectivity. Participants were

asked questions about their access to digital technolo-

gies and resources, aligned with Van Dijk’s model of

digital exclusion:

• What technologies do you use in your day to day

life and how often do you use them?

• What skills do you need to use technology?

• What other barriers stop you using technology on

a daily basis?

• How does digital technology impact your daily

life?

• How would you describe your digital inclusion in

society?

To capture a full portrait of each participant’s digital

exclusion, follow-up questions were asked to identify

speciﬁc barriers that related to digital exclusion.

Between one and four weeks after the interviews,

all seventeen participants took part in one of two face-

to-face focus groups. Both the interviews and focus

group were transcribed verbatim and ’cleaned’ man-

ually to remove personally identiﬁable content. The

qualitative data recorded were collated and then im-

ported into NVivo 14 software (QSR International,

2023). Data were analysed using deductive thematic

analysis, with overarching themes deﬁned by Van

Dijk’s model of digital exclusion (van Dijk, 2020),

and additional themes relating to perceptions of Voice

AI technologies. Participants from Pendeen are in-

dicated by the label ’Pend ’, and St Austell partici-

pants are indicated by the label ’StA ’, followed by

an anonymous ID number.

3 RESULTS

Themes extracted from the interviews and focus

groups have been grouped based on the components

HEALTHINF 2025 - 18th International Conference on Health Informatics

736

of Van Dijk’s model of digital exclusion, focusing

speciﬁcally on the four hierarchies of access and the

associated outcomes. An additional theme relating

to the use of Voice AI systems was included, show-

ing participant’s awareness and previous engagement

with the technology. Subthemes are described in

headers within each theme.

3.1 Attitudes and Motivation

3.1.1 Negative Attitudes to Technology

Some participants perceived technology negatively

due to local services closing, ”I think we’re going too

far. . . . For instance, they are cutting down on fa-

cilities like post ofﬁce, banks” (Pend

6). There were

also concerns over the reliability of technology-based

services, ”I don’t think we should rely on technology

because, yeah, yeah, if it just one switch. It could

all go off” (StA 2). There was also some reluctance

from participants to rely on the internet as a source of

information, and that it was considered less trustwor-

thy than the local community, ”I’d rather go to some-

one who’s actually got the expertise and ask them how

they use it and how they found it. I would trust peo-

ple’s opinion of it rather than the internet’s opinion”

(Pend 8).

In contrast to the negative perceptions of tech-

nology, some participants positively appraised their

devices, ”The mobile phone is a wonderful device”

(Pend 10).

3.1.2 Motivations to Use Technology

Participants expressed a motivation to get more out of

technology to allow them to take part in more digital

activities, ”The more I learn on the digital side, then

hopefully I will get better at doing things and there-

fore be able to do more” (Pend 2). The pervasiveness

of technology was also a motivating factor, to avoid

digital exclusion ”So much is now becoming digital

and I need to get that knowledge up. So I thought,

well, if anything that helps me. The fear is good”

(Pend 2).

3.1.3 Perceived Generational Differences in

Attitudes to Technology

Participants acknowledged not growing up with tech-

nology inﬂuences their perspectives, ”We grew up

with black and white TV, so it’s all ... quite strange

to us” (Pend 2). Participants also suggested that the

skill gap between younger and older adults is insur-

mountable, ”Obviously we’re never going to get to the

stage where kids grown up with it” (Pend 2). Addi-

tionally older adults were characterised as more likely

to be digitally excluded due to poorer digital skills and

device access, ”Older people who aren’t savvy with

even technology or haven’t even got access to it [the

internet]” (StA 2).

3.2 Access to Technology

3.2.1 Physical Access to Technology

Participants reported good access to digital devices,

“You know, we use the emails and computers, any-

thing I need” (Pend 4). Some participants expressed

concerns over the costs associated with acquiring

technologies, “I would hope it would become more

inclusive. That again depends on whether they can

afford the technology . . . Whether the families can

afford the technology” (Pend 1). Participants also ex-

pressed concern about the age of their devices, and

whether they would need replacing over time, “I’m

concerned that the equipment that I’ve got, which was

pretty good two years ago when my husband died will,

in fact eventually need to be replaced. Haven’t got a

clue when that should happen ... I would feel very un-

certain on that and ... I’m not alone in that” (Pend 2).

Participants that worked with vulnerable groups with

limited access to devices stated there was a need to

share hard copies of information to support social in-

clusion, ”I’ll try and print off all the information as

well for ones that haven’t got access to mobile phones

and things” (StA 2).

3.2.2 Material Access to Technology

The largest barrier to digital access for these partici-

pants is the local infrastructure. Participants reported

consistently poor connectivity that was related to the

remote locale, “A lot of us live right away from the

hub of everything ... So I’m outside all the systems,

as are the other three in in our lane” (Pend 8). Con-

nectivity issues may be partly related to the physical

properties of where participants lived, and this issue

may be widespread in Cornwall, ”I mean, even ours

is parlous, our internet. Because we’ve got a granite

house. I think the stone interferes with it. So I think

that’s a problem across Cornwall isn’t it?” (StA 3).

The poor infrastructure was perceived to result

from a lack of concern from internet providers in ru-

ral areas, which resulted in outdated equipment, “It’s

just crazy that in this day and age, everything is com-

ing over land and even electricity cables and every-

thing. And really by now it should be in the ground. If

you go to London or, or any metropolis, they haven’t

got a load of wires, you know, going across. It’s all

Coastal and Rural Digital Exclusion: The Case for Voice AI

737

you know beautifully underground and they have they

have unlimited access to huge bandwidth on their In-

ternet” (Pend 7). Ultimately, participants articulated

the need for the infrastructure to undergo develop-

ment to meet the standards of other areas of the coun-

try, ”Cornwall needs to catch up with the Internet”

(StA 1). The poor connectivity means that people in

the local region faced with poor connectivity are more

likely to rely on landline phones to stay connected,

”You’re talking about areas, coastal areas and areas

where there isn’t necessarily brilliant Internet cover-

age and brilliant mobile signals. We forget there are

people that totally rely on their landlines because they

can’t have the technology” (StA 5).

3.3 Digital Skills

Participants described attempts to learn digital skills

and identiﬁed a clear need for support to learn these

skills, “I’d just like to be with somebody ... who

uses them [digital technologies], and would be able to

show me what the problems are” (Pend 7). Technol-

ogy novelty was reported as being a source of uncer-

tainty, “It’s new and I’m not familiar with. It. And I’m

not so sure what I’m doing” (Pend 10). Participants

reported the need to support vulnerable members of

the community in the use of technology for important

everyday tasks, and that resources to upskill vulnera-

ble individuals are limited, ”We’ve got a lady, haven’t

we in our group and she really struggled. She, you

know, even to pay a bill, you know, something as sim-

ple as that ... we help her obviously. But if she didn’t

have us, you know, what would she do? The help isn’t

out there” (StA

2).

3.4 Usage Inequality

Most participants reported using digital technologies

with some regularity, “I use them, a smartphone and a

laptop. Well, every day” (Pend 3). There were some

participants that used their devices less frequently,

“Mobile phones we, you know, we don’t carry them

around like people do an extension to their arm sort

of thing, but we take them out when we go. But it’s

usually just in case an emergency” (Pend 5).

Participants described a limited number of uses

for different technologies that were largely passive,

including watching media and reading the news, “Get

up the various sites that tells me what’s happening.

Usually the war in Ukraine” (Pend 5). Digital tech-

nologies were also used by some participants to ﬁnd

information, ”If we’re stuck with a problem, the in-

ternet’s the go to place. And often when you’re talk-

ing about phoning places, I’ve almost got to the point

now where you don’t bother with the phone because

usually it’s online” (StA 5).

3.5 Outcomes of Digital Access

The closing down of services in local towns led partic-

ipants caused signiﬁcant difﬁculties for participants,

“Yes, you can go online. Well, number one, as it hap-

pened when my husband died and I had to go in with

quite a lot of paperwork. How do you do that when

they’re not there?” (Pend 6).

The lack of reliable access to community infor-

mation due to the transition to digital platforms was

highlighted as a concern, ”It’s so easy to be out of the

loop, isn’t it? And when you said about relying on

family and friends, that’s what happens now, isn’t it?

... if you haven’t got somebody advocating for you to

do that, it must be the most isolating place” (StA 5).

3.6 Voice AI Use

Participants reported that they were aware of smart

speaker and voice assistant technologies, with some

participants saying that they already used a smart

speaker voice-based interface, ”We use Alexa. Some-

times we use an Alexa like what’s the temperature go-

ing to be today? Things like that” (Pend 8). Partici-

pants also highlighted how Voice AI technologies can

beneﬁt those with mobility impairments, ”It was so I

can just turn the TV on and off and not have to keep

struggling to get up to get to it. You know I couldn’t

get to it. It was really difﬁcult. So in that way, yeah,

it was brilliant (StA 2). The use of smart speakers

was undermined by the poor infrastructure render-

ing them unreliable, ”Like you know, having Alexa in

your house and all of that, I ﬁnd really important. But

the whole thing falls down. Because of our appalling

internet access, and it is appalling” (Pend 8).

Participants also highlighted that there may be

some individuals that may be reluctant to speak with

a Voice AI system due to its artiﬁciality, ”And I know

some people would just, you know, just wouldn’t want

to engage with it at all because it’s not a real per-

son” (StA 3). The use of Voice AI systems should

also be clearly articulated to the user, as the realistic

nature of Voice AI systems may deceive users, ”A lot

of older people, when I think about my mother, like,

would think it’s [the AI] a real person” (StA 3).

4 DISCUSSION

This study describes the characteristics of coastal and

rural digital exclusion faced by two groups of partici-

HEALTHINF 2025 - 18th International Conference on Health Informatics

738

pants co-designing a Voice AI system delivered via a

phone line. The ﬁndings show the complexity of dig-

ital exclusion with varying levels of digital, access,

skills, and usage across participants, with the most

signiﬁcant challenge being the the poor local infras-

tructure. The demographics of the participants also

inﬂuences their digital exclusion, as they reported un-

familiarity with technology due to not growing up

with it, outdated devices and uncertainty as to how

and when to upgrade them (Ueno et al., 2023).

These data highlight how digital exclusion is

a multifaceted spectrum and how factors (Helsper,

2021; van Dijk, 2020) interact to create digital dis-

parities. For example, the largest reported difﬁculty

faced by participants was the unreliable infrastruc-

ture in these coastal and rural locales (Henderson

and Roche, 2020) and how this contrasts to the more

reliable infrastructure in urban areas of the country

(Philip et al., 2017). This limits the utility of digi-

tal devices and reduces access to digital information

and resources (Salemink et al., 2017). These poorer

outcomes in turn negatively impact attitudes towards

technology and the motivation to use digital devices

on a daily basis (Vaportzis et al., 2017), reinforcing

digital exclusion. Consequently, approaches to tack-

ling digital exclusion need to be holistic, and tackle

underlying issues that restrict digital inclusion.

Due to the increasing transition of community

services and resources to online platforms (Spanakis

et al., 2021), digital exclusion can have a negative im-

pact on health and wellbeing, exacerbating existing

health inequalities. Participants described how vul-

nerable people that they know face isolation due to

this as they may have reduced access to technology

and lower digital skills. This social marginalisation

can be exacerbated by the geographic marginalisation

resulting from the poorer infrastructure in coastal and

rural areas of the country compared to the urban areas

(Philip et al., 2017). This isolation can also negatively

impact digital skills for older adults, as previous re-

search has shown that older adult internet users tend

to have greater social support (Friemel, 2016) and that

single-person older-generation households have more

limited technology support and consequently access

to digital resources (Ueno et al., 2023). This means

that digitally excluded older adults may be unable

to access important health and social resources (Ha-

jek and K

onig, 2021), increasing social isolation and

highlighting the importance of identifying approaches

to facilitate access to online information for digitally

excluded members of society (Goldman et al., 2023).

4.1 Study Limitations

The main limitation of this study is that it em-

ployed small, self-selected samples from two loca-

tions. While this does provide valuable, in-depth in-

sight into these speciﬁc contexts, these ﬁndings may

not necessarily be applicable to other rural or coastal

areas or populations within these regions. Further-

more, the sample featured a heavy gender imbalance,

as sixteen of the seventeen participants across both

groups were female. Previous studies of digital exclu-

sion highlight a gender imbalance, as females report

greater levels of digital exclusion than males (Hargit-

tai et al., 2019). This has been reported to be due

to lower levels of digital literacy as well as older fe-

males being perceived as less competent in the use

of technologies (Gallistl et al., 2021). The imbal-

ance observed here may then be due to self selection

bias, as ICONIC is designed to support digital inclu-

sion through the co-design of new technologies, and

may then attract individuals that want to improve their

own digital skills and access. Despite this bias, partic-

ipants still reported varying levels of digital inclusion.

4.2 How Could Voice AI Promote

Digital Inclusion for Coastal and

Rural Communities?

The data reported here suggest that Voice AI systems

could be used to support digital inclusion in coastal

and rural contexts. However, typical access platforms

for Voice AI systems (e.g. smart speakers) are un-

suitable due to infrastructure issues, despite some ac-

ceptance of the technology. The ICONIC project’s

solution, to deliver a Voice AI system via a phone line

(Jones et al., 2024) may be an acceptable alternative

platform. Using a landline phone is familiar and ac-

cessible for these participants, evidenced by it being

the primary platform for interviews, reducing the need

for digital devices or digital skills. The system would

still be dependent on a basic level of phone infrastruc-

ture, i.e. through a landline or mobile signal, how-

ever, it would be more robust to the internet outages

reported by participants. As voice interaction is de-

signed to replicate natural conversation, it represents

a low-tech form of digital interaction, and presents an

alternative to traditional visual interfaces, beneﬁting

individuals with sight impairments (Gu et al., 2020).

Voice commands are more intuitive for older adults

than the potentially complex or unfamiliar UI of digi-

tal devices (Pradhan et al., 2020), and would therefore

be easier for people with few digital skills to use.

Whilst Voice AI has the potential to support access

to digital resources in coastal and rural communities,

Coastal and Rural Digital Exclusion: The Case for Voice AI

739

participants did articulate the potential resistance to

using a Voice AI system due to its artiﬁcial nature.

Part of this resistance may be due to users being un-

aware that they are speaking to an AI. More formative

work needs to be undertaken to negative attitudes to

Voice AI systems. Part of this work should explore

how best to obtain informed consent from users of the

Voice AI system to ensure that they are aware that

they are interacting with an AI, and would be an im-

portant consideration in the system’s design.

5 FUTURE WORK

This ICONIC project (Jones et al., 2024) will run

a series of intergenerational co-design workshops to

create a Voice AI system that can be accessed via

a phone line, exploiting the off-the-shelf availability

of large-language models (Brown et al., 2020) com-

bined with highly accurate end-to-end speech-to-text

systems (Hannun et al., 2014). As participants re-

ported age-related differences in technology use and

skills, similar to previous literature (Vaportzis et al.,

2017), ICONIC’s intergenerational approach to co-

design might support a more equitable design of the

Voice AI technology. Contributions from these differ-

ent perspectives and life experiences may then lead to

a Voice AI technology that is more attuned to the spe-

ciﬁc needs of potential users bases and consequently

be more acceptable when deployed (Thabrew et al.,

2018).

6 CONCLUSIONS

This study highlights the barriers to digital inclu-

sion in rural, coastal communities, particularly among

older adults. Similar to other studies exploring the

digital exclusion of older adults in rural communities,

participants described issues with the coastal and rural

digital infrastructure, as well as with digital skills and

access to devices (Ge et al., 2022; Ueno et al., 2023).

Addressing these infrastructure issues is key to sup-

porting digital inclusion in coastal and rural commu-

nities. However, due to the economic barriers associ-

ated with infrastructure development, it is important

to explore tools such as Voice AI systems, particu-

larly if deployed via a device agnostic platform such

as a phone line. This technology has the potential to

mitigate some of the digital exclusion faced by older

populations in coastal, rural UK communities by by-

passing the limitations of unreliable infrastructure and

the requirement for internet-ready devices and digital

skills. This technology could consequently promote

digital health equity through supporting digitally ex-

cluded populations to access important digital health

and social resources and reduce health inequalities

and loneliness.

ACKNOWLEDGEMENTS

This paper is presented on behalf of the ICONIC

project that includes Marius Varga, Dena Bazazian,

Swen Gaudl, Alejandro Veliz Reyes, Daniel Maudlin,

Chunxu Li, Sheena Asthana, Kerry Howell, Shang-

ming Zhou, Emmanuel Ifeachor, and Hannah Brad-

well as co-applicants and advisors and Lauren Tenn

and Linan Zhang (Media and Administration Ofﬁ-

cers). We thank our participants and partners or-

ganisations. Intergenerational co-creation of novel

technologies to reconnect digitally excluded people

with community & cultural landscapes in coastal

economies (ICONIC) is funded by UK Research and

Innovation Engineering and Physical Sciences Re-

search Council Grant Ref: EP/W024357/1.

REFERENCES

Brown, T., Mann, B., Ryder, N., Subbiah, M., Kaplan, J. D.,

Dhariwal, P., Neelakantan, A., Shyam, P., Sastry, G.,

Askell, A., et al. (2020). Language models are few-

shot learners. Advances in neural information pro-

cessing systems, 33:1877–1901.

Department for Environment Food and Rural Affairs

(2024). Statistical digest of rural england: 1 - pop-

ulation.

Friemel, T. N. (2016). The digital divide has grown old:

Determinants of a digital divide among seniors. New

media & society, 18(2):313–331.

Gallistl, V., Rohner, R., Hengl, L., and Kolland, F. (2021).

Doing digital exclusion–technology practices of older

internet non-users. Journal of Aging Studies, 59.

Ge, B., Harrison, B., Hayes, N., Mason, K., Perez, D.,

Wagg, S., Walker, T., and Wilkes, M. (2022). Digital

poverty transformation: Accessing digital services in

rural northwest communities. In The British Academy,

London.

Goldman, S., Delore, E., Flesh, S., Langer, D., and Mazor-

Karsenty, T. (2023). Digital exclusion and covid-19:

Does communication technology correlate with well-

being? Physical & Occupational Therapy In Geri-

atrics, 41(1):143–158.

Gu, J., Wang, X., Yao, X., and Hu, A. (2020). Under-

standing the inﬂuence of ai voice technology on vi-

sually impaired elders’ psychological well-being: An

affordance perspective. In Human Aspects of IT for

the Aged Population. Technology and Society: 6th In-

ternational Conference, ITAP 2020, Held as Part of

the 22nd HCI International Conference, HCII 2020,

HEALTHINF 2025 - 18th International Conference on Health Informatics

740

Copenhagen, Denmark, July 19–24, 2020, Proceed-

ings, Part III 22, pages 226–240. Springer.

Hajek, A. and K

onig, H.-H. (2021). Social isolation and

loneliness of older adults in times of the covid-19 pan-

demic: can use of online social media sites and video

chats assist in mitigating social isolation and loneli-

ness? Gerontology, 67(1):121–124.

Hannun, A., Case, C., Casper, J., Catanzaro, B., Diamos,

G., Elsen, E., Prenger, R., Satheesh, S., Sengupta,

S., Coates, A., et al. (2014). Deep speech: Scal-

ing up end-to-end speech recognition. arXiv preprint

arXiv:1412.5567.

Hargittai, E., Piper, A. M., and Morris, M. R. (2019). From

internet access to internet skills: digital inequality

among older adults. Universal Access in the Infor-

mation Society, 18:881–890.

Helsper, E. J. (2021). The digital disconnect : the causes

and consequences of digital inequalities / Ellen J.

Helsper. Sage Publications, London.

Helsper, E. J., van Deursen, A. J. A. M., and Eynon, R.

(2015). Tangible outcomes of internet use: from digi-

tal skills to tangible outcomes project report.

Henderson, D. and Roche, N. (2020). Examining the pol-

icy mix for broadband deployment in wales: The role

of informal coordination in the last mile. Local Econ-

omy, 35(1):48–67.

Jones, R. B., Baxter, R., Varga, M., Hagen, O., Aly,

A., Bazazian, D., Veliz Reyes, A., and Gaudl, S.

(2024). Intergenerational technology codesign in de-

prived coastal regions. In ICDS 2024: The Eigh-

teenth International Conference on Digital Society

ICDS 2024, page 7.

Lythreatis, S., Singh, S. K., and El-Kassar, A.-N. (2022).

The digital divide: A review and future research

agenda. Technological Forecasting and Social

Change, 175:121359.

Mascheroni, G. and

Olafsson, K. (2016). The mobile inter-

net: Access, use, opportunities and divides among eu-

ropean children. New Media & Society, 18(8):1657–

1679.

Mee, P., Gussy, M., Huntley, P., Kenny, A., Jarratt, T., Ken-

ward, N., Ward, D., and Vaughan, A. (2024). Digital

exclusion as a barrier to accessing healthcare: a sum-

mary composite indicator and online tool to explore

and quantify local differences in levels of exclusion.

Universal Access in the Information Society, pages 1–

13.

Ministry of Housing Communities and Local Government

(2019). The english indices of deprivation 2019 sta-

tistical release.

Mitchell, J., Shirota, C., and Clanchy, K. (2023). Fac-

tors that inﬂuence the adoption of rehabilitation

technologies: a multi-disciplinary qualitative explo-

ration. Journal of NeuroEngineering and Rehabilita-

tion, 20(1):80.

Money, A., Hall, A., Harris, D., Eost-Telling, C., McDer-

mott, J., and Todd, C. (2024). Barriers to and facil-

itators of older people’s engagement with web-based

services: Qualitative study of adults aged¿ 75 years.

JMIR aging, 7:e46522.

Park, S. and Humphry, J. (2019). Exclusion by design: in-

tersections of social, digital and data exclusion. Infor-

mation, Communication & Society, 22(7):934–953.

Philip, L., Cottrill, C., Farrington, J., Williams, F., and Ash-

more, F. (2017). The digital divide: Patterns, pol-

icy and scenarios for connecting the ‘ﬁnal few’in ru-

ral communities across great britain. Journal of rural

studies, 54:386–398.

Pradhan, A., Lazar, A., and Findlater, L. (2020). Use of in-

telligent voice assistants by older adults with low tech-

nology use. ACM Transactions on Computer-Human

Interaction (TOCHI), 27(4):1–27.

Pradhan, A., Mehta, K., and Findlater, L. (2018). ” accessi-

bility came by accident” use of voice-controlled intel-

ligent personal assistants by people with disabilities.

In Proceedings of the 2018 CHI Conference on human

factors in computing systems, pages 1–13.

Quan-Haase, A., Mo, G. Y., and Wellman, B. (2017). Con-

nected seniors: How older adults in east york ex-

change social support online and ofﬂine. Information,

Communication & Society, 20(7):967–983.

Rural Services Network (2022). House of lords commu-

nications and digital select committee inquiry ‘digital

exclusion and the cost of living’.

Salemink, K., Strijker, D., and Bosworth, G. (2017). Rural

development in the digital age: A systematic literature

review on unequal ict availability, adoption, and use in

rural areas. Journal of Rural Studies, 54:360–371.

Sin, J., Munteanu, C., Waycott, J., N. Brewer, R., Sayago,

S., Lazar, A., and Weber, A. (2022). Alexa, tell me

a joke!:” voice interfaces are truly inclusive”. In CHI

Conference on Human Factors in Computing Systems

Extended Abstracts, pages 1–3.

Song, J. (2023). Advanced Virtual Assistants - A Window

to the Virtual Future, chapter The implications of pro-

viding voice-based chatbots in public service for digi-

tal inclusion and public communication. IntechOpen.

Spanakis, P., Peckham, E., Mathers, A., Shiers, D., and

Gilbody, S. (2021). The digital divide: amplifying

health inequalities for people with severe mental ill-

ness in the time of covid-19. The British Journal of

Psychiatry, 219(4):529–531.

Thabrew, H., Fleming, T., Hetrick, S., and Merry, S. (2018).

Co-design of ehealth interventions with children and

young people. Frontiers in psychiatry, 9:481.

Ueno, A., Dennis, C., and Dafoulas, G. A. (2023). Digi-

tal exclusion and relative digital deprivation: Explor-

ing factors and moderators of internet non-use in the

uk. Technological Forecasting and Social Change,

197:122935.

van Dijk, J. (2020). The Digital Divide. Polity Press.

Vaportzis, E., Giatsi Clausen, M., and Gow, A. J. (2017).

Older adults perceptions of technology and barriers to

interacting with tablet computers: a focus group study.

Frontiers in psychology, 8:294305.

Coastal and Rural Digital Exclusion: The Case for Voice AI

741