Glossary

Every term, plainly explained

Plain-English definitions for AI, hackathons, devices and tech access across the Global South, from the Tembi Labs team.

AI & Machine Learning Devices & Hardware Digital & AI Literacy Global South & Development Hackathons & Building Tech Access & the Divide

Browse A–Z

All terms, A–Z

Every definition has its own page. Jump straight to the term you need.

A B C D E F G H I J K L M N O P Q R S T U V W X Y Z

Jury

Knowledge Transfer

Zero-Rating

AI & Machine Learning

Artificial Intelligence (AI): Software systems that perform tasks normally associated with human intelligence, such as understanding language, recognising images, or making decisions. Modern AI learns patterns from data rather than following fixed, hand-written rules.
Machine Learning: A branch of AI in which a system improves at a task by learning statistical patterns from examples instead of being explicitly programmed. The more relevant data it sees, the better its predictions usually become.
Deep Learning: A form of machine learning that uses neural networks with many layers to learn complex patterns directly from raw data such as text, images, or audio. It powers most of today's language and image models.
Neural Network: A model loosely inspired by the brain, built from layers of connected units that pass signals forward and adjust their strength during training. It is the core building block of deep learning.
Large Language Model (LLM): An AI model trained on very large amounts of text to predict and generate language. LLMs power chatbots, writing tools, and coding assistants, and are the tools students use at Tembi Labs hackathons.
Generative AI: AI that creates new content such as text, images, code, or audio rather than only classifying or predicting. It is the category most students interact with directly.
Prompt: The input text or instruction a person gives an AI model to get a useful response. Clear, specific prompts usually produce better results.
Prompt Engineering: The practice of designing and refining prompts to get reliable, high-quality output from an AI model. It includes giving context, examples, and clear constraints.
Token: The small unit of text, roughly a word fragment, that a language model reads and generates. Model limits and pricing are usually measured in tokens.
Inference: The stage where a trained model is actually used to produce answers or predictions, as opposed to the training stage where it learns. Every chatbot reply is an inference.
Training: The process of teaching a model by exposing it to data and adjusting its internal parameters so its predictions improve. Training large models is computationally expensive.
Fine-Tuning: Adapting an already trained model to a specific task or domain by training it further on a smaller, focused dataset. It is cheaper than training a model from scratch.
Pre-training: The first, large-scale training stage where a model learns general patterns from broad data before being specialised. It gives foundation models their broad capabilities.
Retrieval-Augmented Generation (RAG): A technique that lets a language model look up relevant documents and use them when answering, so responses stay grounded in real, current sources. It reduces made-up answers.
Embedding: A numerical representation of text, images, or other data that captures meaning, so similar items sit close together in a mathematical space. Embeddings power search and recommendation.
Vector Database: A database built to store and search embeddings by similarity rather than exact match. It is a common backbone for RAG and semantic search.
Transformer: The neural network architecture behind most modern language models, which uses an attention mechanism to weigh how words relate to each other. It made today's LLMs possible.
Parameter: An internal value a model adjusts during training to store what it has learned. Large models have billions of parameters, which partly explains their capabilities and cost.
Hallucination: When an AI model produces text that sounds confident but is factually wrong or invented. Recognising and checking for hallucinations is a core AI-literacy skill.
Context Window: The maximum amount of text, measured in tokens, that a model can consider at once when generating a response. Longer context windows let models handle bigger documents.
Multimodal AI: AI that works across several types of input or output, such as text, images, and audio together. A multimodal model can, for example, describe a photo in words.
Computer Vision: The field of AI focused on interpreting images and video, including tasks like object detection and image classification. It underpins applications from medical imaging to agriculture.
Natural Language Processing (NLP): The area of AI concerned with understanding and generating human language. It covers translation, summarisation, sentiment analysis, and question answering.
AI Agent: A system that uses an AI model to plan and take actions across multiple steps, such as searching, calling tools, and updating files, to reach a goal. Agents move AI from answering to doing.
Foundation Model: A large model trained on broad data that can be adapted to many downstream tasks. LLMs and large image models are common examples.
Open-Source Model: An AI model whose weights are publicly available, so anyone can run, study, or adapt it, often for free. Open models lower cost and are important for access in low-resource settings.
Supervised Learning: Training a model on labelled examples, where each input comes with the correct answer. It is used for tasks like classification and prediction.
Unsupervised Learning: Training a model on data without labels, so it finds structure such as clusters or patterns on its own. It is useful when labelled data is scarce.
Reinforcement Learning: A method where an agent learns by trying actions and receiving rewards or penalties, gradually improving its strategy. It is used in robotics, games, and model alignment.
RLHF (Reinforcement Learning from Human Feedback): A technique that fine-tunes models using human preferences about which responses are better, helping them become more helpful and safe. It shaped modern chat assistants.
Dataset: A structured collection of data used to train or evaluate a model. The quality, size, and diversity of a dataset strongly shape what a model learns.
Bias in AI: Systematic unfairness in a model's outputs, often because the training data underrepresents or misrepresents certain groups. Addressing bias matters for AI that serves the Global South.
Chatbot: A program that holds a conversation with users in natural language, today usually powered by a language model. It is the most common way people first meet generative AI.
API (Application Programming Interface): A defined interface that lets one program request services or data from another, such as an app calling an AI model. APIs let builders add AI features without training their own models.

Devices & Hardware

Refurbished Device: A used device that has been tested, repaired, and cleaned to work reliably again, often at a fraction of the cost of new hardware. Tembi Labs builds computer rooms from refurbished enterprise laptops.
Enterprise Laptop: A business-grade laptop built for durability, security, and long service life, which makes it well suited to reuse. Companies retire large fleets of them while they still have years of life left.
Computer Lab: A room equipped with shared computers where students can learn and build, especially valuable where personal devices are scarce. Building and equipping computer rooms is Tembi Labs' longer-term goal.
Device Lifecycle: The full span of a device's use, from purchase through service to retirement and reuse or recycling. Extending the lifecycle keeps working hardware out of the waste stream.
Right to Repair: The principle and growing set of laws that people should be able to repair their own devices, with access to parts, tools, and information. It supports longer device life and less waste.
IT Donation: Giving used computers or other IT equipment to an organisation that can put it to good use. For companies it can combine sustainability, social impact, and tax benefits.
Sachspende: A donation of goods rather than money, such as laptops or equipment. In Germany a Sachspende can, under certain conditions, be tax-deductible for the donor.
Spendenbescheinigung: In Germany, the official donation receipt a recognised charitable organisation issues so a donor can claim the gift on their taxes. For in-kind gifts like laptops it must document the item's value.
Data Wiping: Securely and permanently erasing all data from a device before it is reused or donated, so no personal or company information remains. It is a required step in responsible device reuse.
IT Asset Disposition (ITAD): The managed process of retiring IT equipment safely, including data wiping, refurbishment, resale, donation, or recycling. It lets companies release old hardware responsibly.
E-Waste Recycling: Recovering materials from electronic devices that can no longer be used, so valuable metals are reclaimed and hazardous parts are handled safely. Reuse should come before recycling wherever possible.
Circular Economy: An economic model that keeps products and materials in use as long as possible through reuse, repair, and recycling, instead of a make-use-throw-away line. Refurbishing devices is a clear example.
Total Cost of Ownership (TCO): The full cost of a device over its life, including purchase, energy, maintenance, and disposal, not just the sticker price. Refurbished hardware often lowers TCO for institutions.
Thin Client: A lightweight computer that relies on a central server or the cloud to do most of the work, using little local hardware. It can cut cost and power needs in shared computer rooms.
Chromebook: A low-cost laptop that runs mainly web-based applications, valued for affordability and easy management. It is a common choice for education on tight budgets.
Raspberry Pi: A very small, inexpensive single-board computer widely used for learning, tinkering, and low-cost computing projects. It makes hands-on tech education affordable.
Solar Power: Electricity generated from sunlight, which can power devices and computer rooms where the grid is weak or absent. It links device access to reliable energy.
Off-Grid: Operating without a connection to the public electricity grid, relying on local sources such as solar and batteries. Off-grid setups can bring computing to remote places.
Hardware Sponsorship: When a company funds or supplies devices for a programme, event, or institution. It is one way businesses support Tembi Labs' computer rooms and hackathons.
Device Provisioning: Setting up devices with the software, accounts, and settings needed before handing them to users. Good provisioning makes donated hardware usable from day one.
Operating System: The core software that manages a device's hardware and runs its applications, such as Windows, macOS, Linux, or Android. The choice affects cost, security, and what software will run.
Warranty: A promise from a seller or refurbisher to repair or replace a device if it fails within a set period. A warranty on refurbished hardware gives institutions confidence to adopt it.
Depreciation: The accounting decline in a device's value over time as it ages and is used. It is why companies often retire hardware that still works well.
Corporate Volunteering: When a company enables its employees to give time and skills to social causes, often during work hours. It can pair naturally with device donations and hackathon mentoring.
Corporate Giving: A company's contributions of money, goods, or services to charitable causes as part of its social responsibility. Device donations are a concrete, high-impact form of it.
Tax Deduction for Donations: The reduction in taxable income a donor can claim for eligible gifts to recognised charitable organisations. In Germany it requires a valid Spendenbescheinigung.
Gemeinnützigkeit: In Germany, the legal recognition of an organisation as serving the public benefit, which allows it to issue tax-deductible donation receipts. It is a formal status granted by the tax authorities.
Peripheral: A device connected to a computer to extend its use, such as a monitor, keyboard, mouse, or printer. Complete computer rooms need peripherals as well as the computers themselves.
Server: A computer that provides data, applications, or services to other devices over a network. A single server can support many thin clients or shared workstations.
Networking Equipment: The hardware that connects devices to each other and to the internet, such as routers, switches, and access points. It is essential for turning a set of computers into a working lab.
Grade A/B Refurbished: A grading system for the cosmetic and functional condition of refurbished devices, where Grade A is near-new and lower grades show more wear. It helps buyers know what to expect.
Device Fleet: The full set of devices an organisation owns and manages together. Companies retiring a fleet at once can supply enough units to equip a whole computer room.
Bring Your Own Device (BYOD): A setup where people use their own personal devices for work or study instead of institution-provided ones. Where devices are scarce, BYOD alone leaves many students without access.
Hardware Donation Pipeline: The organised flow that moves retired devices from companies through data wiping and refurbishment to the institutions that receive them. A clear pipeline is what makes device sharing reliable at scale.

Digital & AI Literacy

Digital Literacy: The ability to find, evaluate, use, and create information with digital tools safely and effectively. It is a foundational skill for education, work, and participation in modern society.
AI Literacy: The ability to understand, use, and critically judge artificial intelligence, including its strengths, limits, and risks. Building it is a core aim of Tembi Labs' hackathons and follow-up programmes.
Computational Thinking: A problem-solving approach that breaks problems into clear steps, spots patterns, and designs solutions a computer could follow. It underlies coding but is useful far beyond it.
Upskilling: Learning new skills to do a current role better or keep pace with change, such as adding AI tools to existing work. It helps people stay relevant as technology shifts.
Reskilling: Learning a substantially new set of skills to move into a different role or field. It matters when automation or AI reshapes what jobs exist.
EdTech: Technology designed to support teaching and learning, from learning platforms to educational apps. In low-resource settings, well-designed EdTech can widen access to quality education.
MOOC (Massive Open Online Course): An online course open to large numbers of learners, often free or low-cost, that anyone with internet access can join. MOOCs can bring university-level material to places that lack it.
Mentorship: A relationship where an experienced person guides a less experienced one, sharing knowledge, feedback, and encouragement. Tembi Labs pairs students with mentors during and after hackathons.
Media Literacy: The ability to critically analyse and evaluate media messages, sources, and their intent. It is increasingly tied to spotting AI-generated and misleading content.
Data Literacy: The ability to read, understand, question, and work with data to draw sound conclusions. It helps people make decisions based on evidence rather than assumption.
Coding: Writing instructions that tell a computer what to do, using a programming language. AI tools now lower the barrier to it, but understanding the basics still matters.
STEM Education: Education in science, technology, engineering, and mathematics, seen as key to innovation and skilled work. Widening access to it is central to closing the tech divide.
Lifelong Learning: The ongoing pursuit of skills and knowledge throughout life, not just in formal schooling. It is essential in a world where technology keeps changing.
Blended Learning: An approach that combines in-person teaching with online learning, drawing on the strengths of both. It can stretch scarce teaching resources further.
E-Learning: Learning delivered through digital devices and the internet, from courses to interactive exercises. It can widen access where physical schools or trainers are limited.
Prompt Literacy: The skill of communicating effectively with AI models to get useful, reliable results. It is becoming a practical part of everyday AI literacy.
Critical Thinking: The disciplined skill of analysing information, weighing evidence, and questioning claims before accepting them. It is what keeps AI a tool rather than an unquestioned authority.
Digital Skills Gap: The mismatch between the digital skills people have and those that work and society demand. Closing it is a core purpose of digital-literacy programmes.
Curriculum: The planned set of learning content and goals for a course or programme. A well-designed curriculum makes skills build on each other in a clear order.
Peer Learning: Learning in which people at a similar level teach and support each other rather than relying only on an instructor. Tembi's AI Champions embody this, becoming the people others turn to for help.
Microlearning: Learning delivered in small, focused units that each cover one idea or skill. It fits busy schedules and low-bandwidth settings well.
Certification: A formal credential that shows a person has reached a defined level of skill or knowledge. It can help learners prove ability to employers and universities.
Learning Management System (LMS): Software that hosts courses, tracks progress, and manages learners in one place. It is a common backbone for online and blended education.
Open Educational Resources (OER): Teaching and learning materials that are freely available to use, adapt, and share, often under open licences. They lower the cost of quality education in low-resource settings.
Bootcamp: An intensive, short-term training programme that teaches practical skills, often in coding or tech, in weeks rather than years. It aims to get people job-ready quickly.
Train-the-Trainer: A model where you teach local people to teach others, so knowledge spreads without the original trainer being present. It multiplies impact and builds local capacity.
AI Ethics: The study of how to develop and use AI responsibly, covering fairness, transparency, privacy, and accountability. It shapes whether AI narrows or widens existing inequalities.
Responsible AI: Building and using AI in ways that are fair, safe, transparent, and accountable. Tembi Labs teaches students to build with AI responsibly, not just quickly.
Human-in-the-Loop: A design where a person reviews, guides, or corrects an AI system's output rather than letting it act fully on its own. It keeps human judgement in important decisions.
AI Education: Teaching people how AI works and how to use it well, from basic concepts to hands-on building. It is broader than a single course and spans schools, universities, and workplaces.
Gamification: Using game elements such as points, levels, and challenges to make learning more engaging and motivating. It can raise completion rates in self-paced courses.
Self-Directed Learning: Learning in which people take charge of their own goals, pace, and methods, with less reliance on formal instruction. Access to devices and the internet makes it far more possible.
Learning Outcome: A clear statement of what a learner should know or be able to do after a lesson, course, or programme. Defining outcomes makes it possible to check whether learning worked.

Global South & Development

Global South: A term for the regions of Africa, Latin America, and much of Asia that have historically had less economic and technological power, used instead of older labels like "developing world". Tembi Labs runs its AI hackathons at universities across the Global South.
Global North: A term for the wealthier, more industrialised regions, largely in North America, Europe, and parts of East Asia and Oceania, that concentrate technological and economic power. Most frontier AI models are developed there.
Capacity Building: Strengthening the skills, institutions, and resources a community needs to solve its own problems over the long term. It favours lasting local ability over one-off help.
Brain Drain: The emigration of skilled and educated people from a region, often to wealthier countries, leaving a shortage of local talent. It is a common concern for the Global South.
Brain Gain: The reverse of brain drain, where talent stays, returns, or contributes from abroad, strengthening a region's skills base. Building strong local opportunities helps turn drain into gain.
Diaspora: A community of people living outside their region of origin who often keep strong ties to it. Diaspora networks can channel skills, funding, and knowledge back home. Tembi Labs is led by a diaspora team.
Leapfrog Development: Economic and social progress that skips older stages of development by adopting newer technologies directly, such as mobile banking without traditional banks. AI could enable a new wave of it.
Tech Sovereignty: A region's ability to shape and control the technology it depends on, rather than relying entirely on outside providers. It links access to independence and resilience.
Digital Sovereignty: Control over one's own digital systems, data, and infrastructure, so decisions are not fully dictated by foreign platforms or governments. It is increasingly central to development debates.
Sustainable Development Goals (SDGs): A set of 17 global goals adopted by United Nations members to end poverty, reduce inequality, and protect the planet by 2030. Digital access and education cut across many of them.
Emerging Market: An economy that is growing quickly and industrialising but has not yet reached the wealth and stability of established economies. Many sit in the Global South and hold large young populations.
Frontier Market: A smaller or less developed economy than an emerging market, often with high growth potential but higher risk. Access to technology can be a major growth lever there.
Development Aid: Financial or material help given to lower-income countries to support economic and social progress. Tembi Labs positions its work as talent activation rather than traditional aid.
Localisation: Adapting a product, service, or programme to a specific place, including language, context, and local needs. It is what makes a global tool actually useful on the ground.
Grassroots: Efforts that start from local people and communities rather than being directed from the top down. Grassroots ownership tends to make programmes more durable.
Social Impact: The real, measurable effect an activity has on people's lives and communities. Tembi Labs commits to reporting its impact, budget, and failures openly.
Impact Measurement: The practice of defining, tracking, and reporting the outcomes of a programme with clear indicators. It separates real change from good intentions.
Theory of Change: A clear map of how a programme's activities are expected to lead, step by step, to its intended long-term outcomes. It makes assumptions explicit and testable.
Knowledge Transfer: Passing skills, methods, and know-how from one person or group to another so it takes root locally. Chapter Leads and AI Champions are Tembi's way of keeping knowledge on the ground.
Local Ownership: When the people a programme serves lead and control it, rather than outsiders. It is widely seen as key to programmes that last beyond a single event.
Frugal Innovation: Creating effective, affordable solutions with limited resources, often the norm in the Global South. It shows how constraint can drive, not just limit, creativity.
Informal Economy: Economic activity that is not formally registered or taxed, from street vending to small workshops, which employs a large share of people in many regions. Digital tools can help formalise and grow it.
Remittances: Money sent home by people working abroad, a major source of income for many families and economies in the Global South. It often exceeds development aid in scale.
Microfinance: Small loans and financial services offered to people who lack access to traditional banks, often to start or grow tiny businesses. It is a well-known development tool.
Ecosystem Building: Growing the connected mix of people, skills, funding, and institutions that lets local innovation sustain itself. Tembi Labs frames its work as moving from a first event to a self-sustaining ecosystem.
Tech Hub: A place or organisation that brings together builders, startups, and resources to support technology work, such as a coworking space or incubator. Hubs help concentrate and grow local talent.
AI for Good: The use of artificial intelligence to address social, environmental, and development challenges rather than purely commercial ends. It covers work in health, education, agriculture, and access.
AI in Africa: The growing development and use of artificial intelligence across African countries, including local research labs, startups, and student builders. It reflects a young, fast-growing talent base with rising demand for access.
Talent Activation: Unlocking the potential of people who already have ability by giving them tools, access, and opportunity, rather than treating them as recipients of charity. It is the core idea behind Tembi Labs.
Neocolonialism: The critique that former colonial powers keep economic and cultural control over other regions through indirect means such as trade, technology, and aid. It shapes debates about how development work should be done.
South-South Cooperation: Collaboration and exchange directly between countries of the Global South, sharing skills, trade, and solutions without routing through the Global North. It builds mutual, peer-to-peer strength.
Human Capital: The collective skills, knowledge, and health of a population, seen as a driver of economic growth. Investing in education and access grows a region's human capital.
Inclusive Growth: Economic growth whose benefits reach broad parts of society, not just a narrow elite. Widening technology access is one path toward it.

Hackathons & Building

Hackathon: A time-boxed event, often 24 to 48 hours, where teams build a working solution to a problem from scratch and present it to a jury. The goal is to turn ideas into something real fast.
AI Hackathon: A hackathon where teams build with AI tools such as language models and image generators, so they can ship working prototypes in hours instead of weeks. This is the format Tembi Labs runs at universities across the Global South.
Corporate Hackathon: A hackathon sponsored by a company, usually around a real business problem it wants solved. Tembi Labs runs these as Corporate Challenge Days, where the sponsor sets the problem and keeps the IP.
MVP (Minimum Viable Product): The simplest version of a product that still delivers real value and can be tested with users. Hackathon teams aim for an MVP, not a finished product.
Prototype: An early, working model of an idea built to test whether it makes sense and how it feels to use. Prototypes are meant to be rough and to be thrown away or improved.
Pitch: A short, focused presentation that explains a solution, why it matters, and why it should win. At a hackathon the pitch to the jury often decides the outcome.
Elevator Pitch: A very short summary of an idea, deliverable in under a minute, that captures the core value. It forces builders to say what matters most first.
Jury: The panel that reviews hackathon projects and picks winners based on set criteria such as impact, feasibility, and execution. Tembi juries often include company partners and mentors.
Demo Day: The closing session where teams show their finished projects to an audience, jury, or partners. It marks the transition from building to sharing results.
Sprint: A short, focused period of concentrated work toward a specific goal. A hackathon is essentially one intense sprint.
Ship: To release a working product or feature to real users rather than keeping it in progress. Shipping something usable in limited time is the heart of hackathon culture.
No-Code: Tools that let people build apps and workflows through visual interfaces instead of writing code. They lower the barrier for non-programmers to create working software.
Low-Code: Platforms that speed up building by combining visual tools with small amounts of custom code. They sit between no-code and full programming.
Vibe Coding: Building software mainly by describing what you want to an AI assistant in natural language and iterating on what it generates. It lets people ship working apps with little formal coding experience.
Proof of Concept: A small build that shows an idea can actually work in practice, before investing in a full product. It answers the question "is this even possible?".
User Story: A short description of a feature from the user's point of view, usually in the form "as a user, I want X so that Y". It keeps teams focused on real needs.
Wireframe: A simple, low-detail sketch of a screen or app layout that shows structure and flow before visual design. It helps teams agree on what goes where.
Iteration: One cycle of building, testing, and improving a solution based on what was learned. Good hackathon teams iterate quickly rather than perfecting the first attempt.
Agile: A way of working that delivers in small, frequent increments and adapts to feedback instead of following a fixed long-term plan. Its mindset fits the fast pace of hackathons.
Scrum: A popular agile framework that organises work into short sprints with clear roles and regular check-ins. It is widely used in software teams.
Design Thinking: A problem-solving approach that starts with understanding real users, then rapidly prototypes and tests ideas. It keeps solutions grounded in genuine needs.
Ideation: The early phase of generating and shaping many possible ideas before choosing one to build. Hackathons usually open with focused ideation.
Team Formation: The process at the start of a hackathon where participants group into balanced teams, ideally mixing skills such as building, design, and domain knowledge. Strong teams often decide outcomes.
Mentor: An experienced person who guides teams during a hackathon, giving technical, product, or pitch advice without doing the work for them. Tembi Labs brings mentors to every event.
Facilitator: The person who runs the event, keeps time, sets the rules, and helps participants stay unblocked. A good facilitator keeps a hackathon on track without dominating it.
Deliverable: A concrete output a team is expected to produce, such as a working demo, a slide deck, or a repository link. Clear deliverables make judging fair.
Deadline: The fixed point in time by which work must be finished. The tight deadline is what forces hackathon teams to prioritise ruthlessly.
Version Control: A system that tracks changes to code over time so teams can work together, review history, and undo mistakes. It is essential once more than one person builds.
Git: The most widely used version-control tool, which lets multiple people work on the same codebase in parallel and merge their changes. Most hackathon teams rely on it.
Repository: The storage location for a project's code and its full change history, often hosted on a platform like GitHub. Sharing a repository link is a common hackathon deliverable.
Deployment: The step of putting software onto servers or the web so real users can actually reach and use it. A deployed link is stronger proof than code alone.
Tech Stack: The set of tools, languages, and frameworks a team uses to build a product. Choosing a simple, familiar stack helps hackathon teams move fast under time pressure.
Product-Market Fit: The point where a product clearly solves a real problem for a defined group of users who want it. It is the longer-term goal that the best hackathon ideas move toward.

Tech Access & the Divide

Digital Divide: The gap between people who have reliable access to digital technology and those who do not, along with the differences in skills to use it. It runs between and within countries, and it is central to why Tembi Labs works where it does.
Digital Inclusion: The effort to make sure everyone can access, afford, and confidently use digital technology, including devices, connectivity, and skills. It is the goal on the other side of the digital divide.
Digital Exclusion: The situation of being locked out of digital life because of missing access, affordability, or skills. It can deepen existing social and economic disadvantage.
Connectivity: The ability to reach the internet and digital services, shaped by network coverage, speed, and reliability. Without connectivity, devices and skills deliver little.
Broadband: High-speed internet access capable of carrying data-heavy services such as video and cloud tools. Its availability and price vary sharply across regions.
Last Mile: The final stretch of network that connects the wider internet to a home, school, or campus. It is often the hardest and most expensive part to build in remote areas.
Affordability: Whether devices and internet access cost little enough that people can realistically pay for them. Even where networks exist, high prices keep many offline.
Leapfrogging: When a region skips an older technology stage and adopts a newer one directly, such as going straight to mobile internet without fixed-line phones. It can accelerate development.
E-Waste: Discarded electronic devices, which can pollute and harm health when handled unsafely. The Global South often receives disproportionate volumes of it, which is why responsible device reuse matters.
Data Worker: A person who labels data, trains algorithms, or moderates content that AI systems depend on, often for low pay and without access to the AI they help build. Their work is the invisible backbone of the AI industry.
Mobile-First: Designing digital products for smartphones first, because for many people a phone is their only or main internet device. In much of the Global South, mobile-first is not a choice but a reality.
Internet Penetration: The share of a population that uses the internet, a common measure of digital access. Low penetration signals large unmet potential.
Digital Public Infrastructure: Shared digital systems such as identity, payments, and data exchange that a whole society can build on, much like roads or power grids. Well-designed systems can widen access at scale.
Universal Access: The principle that everyone should be able to reach essential digital services regardless of location or income. It guides many public connectivity policies.
Bandwidth: The amount of data a connection can carry in a given time, which limits how well services like video or large downloads work. Low bandwidth shapes how tools must be designed for many regions.
Offline-First: Designing software so it still works without a stable internet connection and syncs when a connection returns. It is important where connectivity is patchy.
Data Poverty: Being unable to afford enough mobile data to participate fully in digital life. It forces hard trade-offs between connectivity and other basic needs.
Access Gap: The measurable difference in who can reach and use technology, broken down by income, location, gender, or age. Naming the gap precisely helps target where to act.
Device Divide: The gap between people who own capable computers and those who rely only on shared or basic devices such as low-end phones. Writing a thesis on a phone is a common symptom.
Meaningful Connectivity: A standard that goes beyond being online at all, requiring a fast enough connection, an adequate device, and enough data to use the internet productively. It reframes access as quality, not just presence.
Digital Rights: Human rights as they apply online, including access, privacy, free expression, and protection from surveillance. They frame access as a matter of fairness, not just infrastructure.
Net Neutrality: The principle that internet providers should treat all traffic equally rather than favouring or blocking particular services. It affects what an open internet looks like in practice.
Zero-Rating: When a provider lets users reach certain apps without it counting against their data allowance. It can widen access but also steer people toward a narrow slice of the internet.
Community Network: A local internet network built and run by the community it serves, often where commercial providers do not reach. It is a grassroots answer to connectivity gaps.
Rural Connectivity: Internet access in sparsely populated areas, where distance and low density make networks costly to build. Closing rural gaps is one of the hardest parts of digital inclusion.
Smartphone Dependency: Relying on a phone as the only computing device, which limits tasks like coding, long-form writing, or multitasking. It quietly caps what talented people can build.
Data Colonialism: The critique that value extracted from people's data flows mainly to a few dominant firms and regions, echoing older colonial patterns. It raises who benefits from the data economy.
Digital Public Goods: Open-source software, open data, and open content that are freely available and serve the public interest. They can lower the cost of building solutions in low-resource settings.
Universal Service: A policy obligation to make basic communication services available to everyone at a reasonable price, including hard-to-reach areas. It is a common tool for narrowing access gaps.
Access to Electricity: Reliable power, without which devices and networks cannot function. In many regions, electricity gaps are the first barrier to digital access, before connectivity or devices.
Tech Divide: A broad term for the imbalance in who develops, owns, and benefits from technology, spanning devices, connectivity, skills, and the AI economy. It is the framing Tembi Labs uses for the North-South gap.
Second-Level Digital Divide: The gap that remains even among people who are online, driven by differences in skills and how effectively they use technology. Access alone does not guarantee opportunity.
Gender Digital Divide: The gap in access to and use of technology between women and men, often wider in lower-income regions. Closing it is a stated priority in many digital-inclusion efforts.