The Digital Divide: Why Technology Access Remains Unequal

The digital divide is often described as the gap between people who have internet access and people who do not. That definition is increasingly inadequate. A mobile signal reaching a village does not guarantee that households can afford data, that the connection is fast enough for online classes or telemedicine, that people own suitable devices, that services work in languages they understand or that users have the skills and confidence to navigate them safely.

Technology access therefore works more like a chain. Infrastructure must exist. Service has to be reliable and affordable. A suitable device must be available when it is needed. Electricity must keep that device and the network functioning. Platforms have to be accessible to people with disabilities. Users need enough digital skill to complete meaningful tasks and enough trust that scams, surveillance or account loss do not make participation too risky. If any important link fails, formal connectivity may exist while practical participation remains limited.

The International Telecommunication Union's latest global framework captures this shift. Its Global Connectivity Report 2025 defines universal and meaningful connectivity through six interdependent dimensions: quality, availability, affordability, devices, skills and security. ITU estimates that around 6 billion people—almost three-quarters of the world's population—were using the internet in 2025, but approximately 2.2 billion remained offline. The remaining divide is increasingly about not only whether people can connect, but what kind of connection they can actually use and what that connection allows them to do. (itu.int)

That distinction matters because education, employment, banking, communication, healthcare and government services increasingly assume digital participation. As technology becomes integrated into ordinary institutions, unequal access stops being only a telecommunications problem. It becomes a question of who can participate fully in modern social and economic life.

Connectivity Has Expanded Dramatically, but Quality Still Follows Wealth and Geography

The global infrastructure gap has narrowed substantially. ITU reports that more than 96% of the world's population is now covered by a mobile broadband network, and 4G reaches around 93% of people globally. Yet those global averages conceal large differences in the quality of connection available to different populations. (itu.int)

The contrast is particularly visible in newer networks. In 2025, 5G covered approximately 55% of the global population, but coverage reached 84% in high-income countries and only 4% in low-income countries. The rural divide remains substantial as well. Globally, 66% of urban residents were within 5G coverage compared with 40% of rural residents, while in low-income countries only 38% of rural residents were covered even by 4G. (itu.int)

These differences matter because internet access is not technologically interchangeable. A weak mobile connection may be sufficient for text messaging while failing during a video consultation, software update, online examination or live class. A worker relying on cloud-based professional software needs something different from someone sending occasional messages. A school with unreliable bandwidth cannot use digital learning in the same way as one with high-capacity fixed broadband.

Infrastructure economics helps explain why the pattern persists. Fibre routes, mobile towers, spectrum, international links, data centres and reliable electricity all require investment. Commercial providers have stronger incentives to invest where population density and purchasing power create faster returns. Remote, sparsely populated, mountainous, conflict-affected or very poor areas are more expensive to connect for each potential customer.

Public policy can alter those incentives through infrastructure sharing, universal-service obligations, subsidies, spectrum policy and direct public investment. But announcing network coverage is not the same as delivering useful connectivity. Programmes need measurable standards for reliability, speed and affordability rather than treating the appearance of a signal bar as the end of the problem.

Electricity belongs inside the same infrastructure discussion. Devices require charging, mobile towers need power and broadband equipment cannot remain reliable through repeated outages without backup systems. Digital access therefore depends partly on energy policy. A school can receive laptops and connectivity equipment while still being unable to use them consistently because electricity is unreliable.

Coverage is the first link in the chain, not the final measure of inclusion.

Affordability and Devices Determine Whether Networks Become Usable

Infrastructure becomes meaningful only when people can afford to use it.

ITU reports that internet services have generally become cheaper, but affordability remains a major barrier. In 2025, around one-third of economies had still not reached the Broadband Commission's affordability target, and ITU reported mobile broadband as unaffordable for many lower-income populations even where the national average price appeared manageable. (itu.int)

Absolute price alone can be misleading. A data package costing the equivalent of a few dollars may represent a minor expense to a wealthy household and a significant share of disposable income to a poorer one. Connectivity competes with rent, food, transportation, electricity, medicines and education. The relevant question is therefore not simply how many dollars a broadband package costs, but how much household purchasing power it consumes.

Within-country inequality can be as important as differences between countries. ITU's 2025 work combining connectivity prices with income-distribution data shows that the same broadband basket can represent a much heavier burden for households in the poorest part of the income distribution than for someone earning the national average. (itu.int)

Device costs create a second threshold. A low-cost smartphone can provide extraordinary capability: messaging, digital payments, maps, video, government services and web access. But a smartphone is not an equivalent substitute for every task. Writing a long report, programming, analysing spreadsheets, producing professional designs or completing a complex application can be considerably harder on a small screen.

Meaningful device access therefore depends on task suitability.

Counting one smartphone as proof that a household is fully digitally equipped can hide important inequalities. Several family members may share that one device. Children may compete for it during online study. One adult may control when other household members can use it. Sensitive searches involving health, finances or relationships may be impossible to conduct privately.

Financial systems influence device access as well. Higher-quality laptops and phones are frequently purchased through instalments, bank cards or formal credit. People outside those systems may need to pay the full cost upfront, rely on older second-hand equipment or continue using devices beyond the point where batteries, storage or operating systems reliably support essential applications.

Repairability affects long-term affordability too. A cracked screen, failed battery or damaged charging port can effectively remove someone from digital participation when repairs cost too much or replacement parts are unavailable. A programme that distributes devices but provides no maintenance, repair or replacement pathway may produce impressive launch numbers followed by declining real access.

The same applies to public computer centres, school labs and donated equipment. Technology projects fail when budgets cover installation but not staff, electricity, repairs, cybersecurity, software licences and eventual replacement.

Durable access is more expensive—and more useful—than one-time distribution.

Skills, Education and Language Determine What People Can Actually Do Online

Owning a connected device does not automatically create digital capability.

Basic digital literacy can involve operating a device, managing passwords, installing software, recognising fraudulent links, uploading documents and navigating common applications. More advanced participation may require spreadsheets, professional platforms, online research, privacy management, digital content creation or the ability to evaluate the credibility of information.

ITU's 2025 evidence suggests a hierarchy in digital skills. Communication-related abilities are relatively common among internet users, while more advanced skills involving problem-solving, safety and content creation develop less evenly. (itu.int)

This produces an important distinction between internet use and effective internet use.

Someone may use messaging applications every day but struggle to submit a government form, recover a compromised account or evaluate whether an online employment offer is fraudulent. Another person may be technically connected but unable to use the professional applications increasingly required in the labour market.

Education strongly influences this divide. Literacy, numeracy, language ability and previous exposure to technology affect how easily people acquire digital skills. Workers employed in technology-intensive occupations continue learning through repeated exposure, while workers in jobs with few digital tools may receive less opportunity to develop those abilities over time. Two people with identical home internet connections can therefore move further apart in digital capability during their careers.

This creates a feedback loop. Educational disadvantage can reduce digital capability, while digital disadvantage can then restrict access to online learning, employment information and educational resources. UNICEF's Digital Education Strategy 2025–2030 explicitly treats connectivity, device access, digital skills and social barriers as interconnected components of educational inequality rather than assuming that placing technology in schools automatically improves learning. (unicef.org)

Language is another form of infrastructure.

A service can technically be available worldwide while being practically inaccessible to someone who cannot understand its instructions, customer support or legal terms. Automated translation and speech technology can reduce some language barriers, but performance remains uneven across languages, dialects and accents. Smaller language communities also have fewer digital resources from which contemporary AI and search systems can learn.

Local content therefore matters. People are more likely to gain value from connectivity when they can find information about their own schools, laws, markets, healthcare systems, prices and public institutions in languages they understand.

Meaningful access is not simply the ability to reach the global internet.

It is the ability to find something useful when you arrive.

Technology Access Is Unequal Within Households as Well as Between Them

National and household statistics can conceal inequality inside families.

A household may be counted as connected because one member owns a smartphone or maintains a data subscription. That does not establish that every adult and child has equal control over the device, enough time to use it or the privacy required for personal activity.

Gender demonstrates this problem particularly clearly.

The GSMA Mobile Gender Gap Report 2026 estimates that women in low- and middle-income countries remained 12% less likely than men to use mobile internet in 2025, representing approximately 200 million fewer women using mobile internet than men. Around 810 million women in those countries were still not using mobile internet. GSMA identifies continuing barriers involving affordability, digital skills and online safety, among others. (gsma.com)

The important word here is control.

A woman may technically live in a connected household while needing permission to use a shared device, facing monitoring of online activity or receiving lower priority when the household decides whose education or employment justifies buying a device. Safety concerns can also operate differently: harassment, image abuse, stalking or financial scams can make online participation more costly for some users.

Age produces another pattern. Older adults may have had far fewer opportunities to learn digital systems through school or work. Complex authentication systems, constantly changing interfaces and terminology written for experienced users can make ordinary online tasks disproportionately difficult. Age itself does not make technological participation impossible; design and training determine much of the difficulty.

This is why assisted access remains valuable. Libraries, community centres, banks, schools and government offices can provide human support to people who need help recovering accounts, completing online forms or understanding unfamiliar systems. Digital inclusion should not assume that the only successful user is someone capable of navigating every platform independently.

The same principle applies to children, migrants, people with low literacy and anyone temporarily unfamiliar with a system.

Human assistance is not evidence that digitalisation has failed.

It can be part of making digital systems genuinely usable.

Disability Shows Why Connectivity and Accessibility Are Different

A person can have an excellent broadband connection, a modern computer and advanced digital skills and still be excluded by inaccessible design.

A blind user may encounter a website whose buttons have no meaningful labels for a screen reader. A deaf person may be unable to use an essential instructional video without captions. Someone with limited dexterity may be blocked by an interface requiring precise mouse movements. A person with cognitive or intellectual disability may face unnecessarily complex language and confusing navigation.

The UN Convention on the Rights of Persons with Disabilities treats access to information and communication technologies as part of the broader right to accessibility. Article 9 covers not only buildings and transport but information, communications and electronic services. (un.org)

This distinction matters because accessibility is frequently treated as an optional feature added after the main technology has already been designed.

Retrofitting is usually harder.

Inclusive design instead asks at the beginning whether the service can be used through different input and output methods, whether text has sufficient structure, whether video has captions, whether forms work with assistive technologies and whether users can request an accommodation when universal design does not meet a particular need.

UNICEF's work on digital inclusion for girls and young women with disabilities makes the same point: gender and disability barriers can compound one another, and accessible digital environments work best when people with disabilities participate in their design. (unicef.org)

Artificial intelligence creates both possibilities and risks in this area.

Automatic captioning, image description, speech interfaces and adaptive communication tools can make information substantially more accessible. But AI systems can also fail disproportionately when disabled users are poorly represented in training or testing. Voice systems may perform badly for atypical speech. Automated hiring tools may treat disability-related employment patterns as negative signals. Authentication systems can be inaccessible to people who cannot complete the assumed biometric or interaction.

Digital inclusion therefore requires accessibility to be treated as technical quality, not a charitable addition.

A platform that excludes part of its intended population is not fully functioning.

Digital Government Can Turn a Technology Gap Into a Rights Gap

Digitalisation becomes especially consequential when governments move essential services online.

Tax filing, welfare applications, identity systems, licences, health appointments and benefit payments can become faster and less costly when designed well. Digital systems may reduce travel, shorten queues and make services available outside office hours.

But once a digital channel becomes compulsory, technology access stops being merely a consumer preference.

A person who lacks a device, data connection, formal identification, digital skill or accessible interface may be prevented from exercising a legal entitlement.

This is why digital-by-default systems need more than functioning software. They need assisted channels, accessible design, realistic recovery processes and alternatives for people who cannot use the primary digital route. Technology should simplify access to rights rather than create a new qualification for receiving them.

Hidden prerequisites deserve particular scrutiny. An online service may assume that every user has a mobile number, permanent address, government identification, bank account or biometric record. People without one of these prerequisites can be excluded before they even encounter the main service.

Account recovery is another underestimated problem.

A service can be easy to join but effectively inaccessible if losing a SIM card or forgetting a password creates a recovery process requiring documents or devices the user no longer possesses. Local technical support therefore belongs inside meaningful access. Users need somewhere to repair devices, resolve account problems and understand unfamiliar procedures.

Private platforms create similar issues when they become infrastructure-like. Messaging platforms, payment systems, app stores, online marketplaces and maps can become so central to economic and social participation that account suspension, pricing decisions or accessibility failures have effects far beyond an ordinary consumer product.

Technology access therefore raises questions not only about infrastructure but about governance.

Who controls the system? What happens when access is withdrawn? Can decisions be appealed? What data must users surrender? Is there an alternative?

As more of ordinary life moves online, these questions become part of digital inclusion.

Trust, Safety and Security Are Part of Meaningful Access

People do not necessarily use a digital service simply because it is available.

They may avoid online banking because scams are widespread. They may distrust a government portal if they do not understand how personal information is stored. Parents may restrict children's internet use because of harassment or exposure to harmful content. A worker may hesitate to upload identity documents through a service whose legitimacy is difficult to verify.

ITU therefore includes security as one of the six dimensions of universal and meaningful connectivity. Its Global Connectivity Report 2025 notes that expanding access also creates exposure to cybercrime, online violence, misinformation and other risks. (itu.int)

This is important because safety and inclusion can sometimes be treated as competing goals.

They should reinforce each other.

Strong consumer protection, clear fraud-reporting systems, effective cybersecurity and digital literacy make people more willing to use services confidently. Poor security creates another hidden cost of participation: users must accept a level of financial or privacy risk that they may reasonably find unacceptable.

Safety barriers can also be unevenly distributed. Women, children, public figures and minority groups may experience different kinds of online harassment. People with limited digital literacy can be targeted by scams built around unfamiliar interfaces or urgent requests for payment.

The correct policy objective is therefore not simply more internet use.

It is safe, informed and useful internet use.

Connectivity that exposes inexperienced users to preventable harm without providing meaningful protection is an incomplete form of inclusion.

Markets, Regulation and Public Investment Shape Who Gets Connected

The digital divide is sometimes framed as though it results mainly from individuals failing to adopt technology.

That overlooks how strongly access is shaped by markets and regulation.

Telecommunications markets with weak competition may sustain higher prices or provide fewer incentives to improve service. Infrastructure-sharing rules can affect whether operators duplicate expensive networks or share towers and fibre. Spectrum policy can influence coverage and capacity. Universal-service funds can subsidise deployment in places where commercial returns are weak.

ITU's Global Connectivity Report specifically identifies regulation, competition and cross-sector coordination as important tools for improving affordability and expanding meaningful access. (itu.int)

Procurement also matters.

Governments, schools and large public institutions purchase huge quantities of software, devices and connectivity. If accessibility, interoperability, security and long-term support are required in procurement contracts, suppliers have incentives to build those characteristics into products from the beginning.

The opposite approach can lock in exclusion.

A government may purchase an educational platform before checking whether it works on inexpensive devices, supports local languages or can be used by students with disabilities. Once data, training and workflows are built around that platform, changing systems becomes expensive. A design mistake that could have been corrected during procurement becomes institutional infrastructure.

Universal-service policy faces a similar problem. Subsidising network construction is useful only if households can subsequently afford subscriptions and devices. A tower built in a low-income area can technically eliminate a coverage gap while leaving actual usage low because data remains unaffordable.

Policy therefore has to follow the chain.

Infrastructure support without affordability policy is incomplete. Device subsidies without digital skills can underperform. Digital-skills programmes cannot overcome inaccessible services. Cybersecurity education cannot compensate for unsafe platform design.

Digital inclusion requires coordination among telecommunications regulators, education systems, disability policy, competition authorities, consumer-protection agencies, local government and social-protection systems.

No single ministry controls all the barriers.

The Best Measure of Technology Access Is What People Can Actually Do

Simple connectivity indicators remain useful. Governments need to know which areas lack broadband coverage, how much data costs and how many households own computers or smartphones.

But those statistics should not be mistaken for the objective itself.

Meaningful technology access is better measured through capability.

Can a student reliably attend an online class and complete the required assignments? Can a worker search and apply for employment? Can a small business receive digital payments? Can a patient book an appointment or use telemedicine? Can a citizen access government benefits? Can a blind person complete the same online form as everyone else? Can users protect accounts and recover them when something goes wrong?

These outcomes reveal barriers that ownership statistics can miss.

A family can own a smartphone while children cannot use it during school hours. A village can have 4G while data prices make video learning unrealistic. A government can provide an online service that people with disabilities cannot navigate. A worker can have broadband but lack the digital skill expected in better-paid jobs.

Access also has to be durable.

Temporary connectivity funded by a pilot project does not create lasting digital participation if service disappears once the subsidy ends. A donated laptop does not produce durable capability when the battery fails and no replacement is available. A one-time computer course does not prepare someone indefinitely for changing cybersecurity threats, AI tools and workplace platforms.

Technology policy therefore needs upgrade paths.

Today's minimum device may not run tomorrow's essential software. Today's network standard may become inadequate as services become more demanding. Today's digital skills will need updating.

The policy target should not be permanent possession of one particular technology.

It should be the ability to continue participating as technology changes.

The Digital Divide Persists Because Several Inequalities Reinforce One Another

Technology inequality cannot be explained by one missing cable or one unaffordable phone.

Geography affects infrastructure costs. Income affects whether households can buy devices and recurring data. Education affects digital skill. Gender can affect ownership and control within households. Disability interacts with design. Language affects usability. Financial exclusion influences device purchasing. Housing conditions affect the ability to study or work online. Weak consumer protection reduces trust. Poorly designed digital government can turn all of these disadvantages into barriers to public services.

These inequalities reinforce one another.

A child with weak educational opportunities may develop fewer digital skills, which can restrict later employment opportunities, which in turn reduces the income available for better devices and connectivity. A rural household may have lower income and weaker broadband at the same time. A disabled woman can encounter gender, affordability and accessibility barriers simultaneously rather than sequentially.

This is why the idea of a technology-access chain is useful.

Coverage matters, but coverage alone is not access.

Affordability matters, but a cheap connection on an unsuitable device can still restrict participation.

Devices matter, but technology that a person cannot operate safely or accessibly does not create capability.

Skills matter, but skills cannot compensate for a network that does not work.

And digital services themselves must be useful enough that people have a reason to remain connected.

ITU's shift toward universal and meaningful connectivity reflects this more mature understanding. The objective is no longer simply to place everyone within reach of a telecommunications network. It is to make it possible for people to use digital technology with sufficient quality, affordability, devices, skills and security to participate meaningfully in modern life. (itu.int)

That requires infrastructure investment.

It also requires competition, education, accessible design, consumer protection, local technical support, inclusive procurement, lifelong digital learning and public services that do not abandon people who need assistance.

The digital divide will not disappear merely because the last village receives a mobile signal.

It narrows when people can turn connectivity into real capability: to learn, work, communicate, create, access services, protect themselves and participate on terms that remain viable as technology continues to change.