Unit 2: Emerging technologies and future of the discipline

I. Orientation

Social sciences study human behaviour, institutions, relationships, cultures, and systems of power. Emerging technologies are changing how these subjects are observed, interpreted, and applied. The central principle is human-centred, evidence-based, and ethically responsible innovation: technology should strengthen social understanding without reducing people to data points or ignoring rights, context, and inequality.

• Interdisciplinary character: Sociology, anthropology, psychology, economics, political science, communication, data science, and computer science increasingly overlap.
• Digital mediation: Social interaction now occurs through platforms, mobile devices, sensors, online communities, and digital public spaces.
• Data pluralism: Evidence may include interviews and observation alongside administrative records, social-media data, geospatial information, images, audio, and sensor data.
• Reflexivity: Researchers must examine how their own concepts, software, sampling choices, and institutional positions shape findings.
• Human accountability: Automated outputs are aids to judgement, not substitutes for ethical reasoning, democratic oversight, or lived experience.
• Core tension: Digital technologies can expand access and accuracy while also producing surveillance, bias, exclusion, misinformation, and unequal control over data.

II. Emerging Technologies in Social Sciences — New tools for social inquiry

Emerging technologies are developing tools whose social applications are still changing. In social sciences, their value lies not simply in novelty but in whether they provide valid, inclusive, and interpretable evidence.

A. Emerging Technologies in Social Sciences

This topic examines technologies that transform data collection, analysis, communication, and social intervention.

• Artificial intelligence: Machine-learning systems identify patterns in text, images, speech, or behaviour. A classifier might label thousands of news articles by topic, but its categories reflect training data and design choices.
• Big-data infrastructure: Cloud storage and distributed computing allow analysis of datasets containing millions of records, such as anonymised mobility traces or online interactions.
• Internet of Things: Connected devices generate continuous information. Wearable activity trackers, for example, can measure steps or heart rate, but access may exclude people without suitable devices.
• Geospatial technologies: Geographic information systems combine location with social variables. Mapping clinic locations against travel time can reveal unequal healthcare access.
• Virtual and augmented reality: Immersive environments support studies of presence, empathy, learning, and spatial behaviour. Results may differ from ordinary life because the setting is simulated.
• Blockchain and distributed ledgers: These systems record transactions across a network rather than in one central database. They may improve traceability, although permanence can conflict with privacy and data deletion.
• Digital platforms: Social-media and collaboration platforms create both research sites and data sources; their algorithms also influence what users see and how they interact.

B. Applications and limitations

Technology becomes socially useful only when its benefits are evaluated against validity, accessibility, and harm.

• Research applications: Digital tools support real-time disaster mapping, online ethnography, automated transcription, network analysis, and participatory planning.
• Validity limitation: A large dataset is not automatically representative. Platform users may differ from non-users by age, income, language, or location.
• Access limitation: Broadband, device ownership, disability, and digital literacy create a digital divide.
• Interpretive requirement: Correlation in a platform dataset does not establish causation; social meaning still requires theory and contextual analysis.

III. Artificial Intelligence and Digital Transformation in Social Sciences — Intelligent systems and institutional change

Artificial intelligence (AI) refers to computational systems that perform tasks associated with human intelligence, including prediction, classification, language processing, and pattern recognition. Digital transformation is broader: it is the organisational and cultural restructuring that occurs when digital systems become embedded in research, education, governance, and professional practice.

A. Artificial Intelligence and Digital Transformation in Social Sciences

The topic concerns both AI techniques and the institutional changes produced by their adoption.

• Supervised learning: A model learns from labelled examples. If 10,000 survey responses are coded as “support” or “oppose,” the model estimates labels for new responses.
• Unsupervised learning: The system finds structures without pre-labelled outcomes. Clustering may group respondents by similar attitudes, but researchers must interpret whether the groups are meaningful.
• Natural-language processing: Tokenisation, sentiment analysis, topic modelling, and large language models process written or spoken language. A sentiment score can overlook irony, dialect, or culturally specific expression.
• Predictive analytics: A model estimates an outcome such as voter turnout or service demand. Prediction is not explanation: a high-performing model may not reveal the mechanism causing the outcome.
• Digital transformation: Universities and public agencies adopt learning-management systems, digital records, automated workflows, dashboards, and remote collaboration. This changes roles, decision speed, accountability, and required skills.
• Human–AI collaboration: Researchers may use AI to transcribe interviews or suggest codes, while humans verify accuracy, preserve context, and make final interpretive decisions.

B. Applications and limitations

AI can increase analytical capacity, but transformation must preserve professional judgement and institutional responsibility.

• Efficiency gain: Automated transcription can convert recorded interviews into searchable text, reducing routine labour while still requiring checks for names, accents, and overlapping speech.
• Bias risk: If historical policing data over-represent certain neighbourhoods, a predictive system may reproduce that unequal attention.
• Opacity problem: Complex models may be difficult to explain to participants or policymakers; interpretable features and documentation are therefore important.
• Organisational impact: Automation may redistribute work rather than eliminate it, creating new duties in data governance, model auditing, and digital security.

IV. Digital Research Methods — Methods for studying digitally mediated life

Digital research methods use digital technologies to collect, analyse, visualise, or communicate evidence. They extend traditional methods rather than making them unnecessary; the method must remain appropriate to the research question and population.

A. Digital Research Methods

The purpose of these methods is to investigate online and offline social processes through systematic, transparent procedures.

• Online surveys: Web questionnaires reach geographically dispersed participants quickly. Randomisation of question order can reduce order effects, but open links may attract self-selected respondents.
• Digital interviews and focus groups: Video calls enable remote participation and recording. Researchers must account for connectivity, private space, camera refusal, and platform fatigue.
• Digital ethnography: Researchers observe online communities, forums, or gaming spaces over time. Participation, field notes, and attention to community norms are central.
• Social-network analysis: Nodes represent actors and edges represent relationships or interactions. Degree centrality counts direct connections; a highly connected account may be influential, although connection quantity is not identical to social power.
• Text and content analysis: Software can count terms, identify themes, or compare discourse across periods. Human coding helps establish conceptual validity.
• Geospatial analysis: Geographic coordinates are combined with social indicators to examine segregation, service access, migration, or environmental inequality.
• Digital trace analysis: Clicks, timestamps, searches, and mobility records reveal patterns of behaviour, but traces are generated by platform design and do not always represent intention.

B. Applications and limitations

Digital methods require methodological fit, data quality checks, and protection of participants.

• Sampling: Researchers should define the population and recruitment frame before collecting data; a public hashtag is not equivalent to all public opinion.
• Triangulation: Combining an online survey with interviews or administrative statistics can test whether different forms of evidence converge.
• Reproducibility: A study should record search dates, software versions, coding rules, and data-cleaning decisions so that analysis can be audited.
• Privacy: Removing usernames may not fully anonymise a dataset because distinctive quotations or location combinations can permit re-identification.

V. Social Innovation and Sustainable Development — Social change through responsible solutions

Social innovation means developing and implementing new ideas, practices, services, or organisational forms that address social needs and improve collective wellbeing. Sustainable development links present welfare with long-term environmental, economic, and social responsibility.

A. Social Innovation and Sustainable Development

This topic connects technological creativity with participation, equity, and the interdependence of social and ecological systems.

• Needs-based innovation: A solution begins with a defined social problem, such as food insecurity or inaccessible transport, rather than with technology alone.
• Co-design: Communities, practitioners, governments, and researchers jointly define problems and test responses. Local knowledge can reveal barriers missed by technical experts.
• Sustainable development dimensions: Social inclusion, economic viability, and environmental protection must be considered together. A low-cost digital service is not sustainable if it generates excessive energy use or excludes older users.
• Circular approaches: Repair, reuse, sharing, and recycling reduce material waste. E-waste illustrates why digital progress must include responsible production and disposal.
• Impact measurement: Indicators may include service reach, reduced emissions, improved wellbeing, employment, or participation. A single output, such as app downloads, does not prove social impact.
• Scalability and context: A successful community project may need adaptation before expansion because infrastructure, culture, and institutional capacity differ across locations.

B. Applications and limitations

Social innovation is effective when communities retain agency and outcomes are assessed over time.

• Application: A participatory mapping project can identify flood-prone roads and help residents and local authorities prioritise evacuation routes.
• Equity test: Benefits should be compared across gender, disability, income, rural–urban location, and other relevant groups.
• Unintended effects: A smart-city camera network may improve traffic management while increasing surveillance of public space.
• Long-term viability: Projects need maintenance funding, local skills, accessible design, and governance arrangements after pilot funding ends.

VI. Future Directions and Professional Competencies — Preparing for evolving practice

Future social-science work will combine substantive social knowledge with digital, analytical, communicative, and ethical competence. Professionals will increasingly interpret complex evidence for institutions and communities rather than merely operate software.

A. Future Directions and Professional Competencies

The future direction of the discipline depends on adaptable professionals who can connect technical capability with social purpose.

• Data literacy: Professionals should understand variables, sampling, missing data, uncertainty, correlation, and visualisation. A percentage without its denominator can mislead interpretation.
• Computational literacy: Basic competence in spreadsheets, databases, coding, or statistical software supports transparent analysis. Version control and documented workflows improve reproducibility.
• AI literacy: Users should understand training data, model evaluation, bias, hallucination, privacy, and limits of automated output.
• Qualitative and interpretive skill: Interviews, observation, discourse analysis, and contextual reasoning remain essential for understanding meaning and experience.
• Interdisciplinary teamwork: Effective projects require communication among researchers, designers, engineers, policymakers, legal specialists, and affected communities.
• Communication competence: Findings must be translated into clear reports, data visualisations, policy briefs, and accessible public explanations without overstating certainty.
• Lifelong learning: Tools change rapidly, so professionals need continuing development in software, regulation, cybersecurity, and research ethics.
• Civic and global competence: Future practice must address platform power, climate change, migration, inequality, digital rights, and culturally diverse knowledge systems.

B. Applications and limitations

Professional competence is demonstrated through responsible decisions, not through technical skill alone.

• Practical standard: A researcher should be able to explain why a tool was selected, what data it uses, how error is measured, and who may be harmed.
• Collaborative standard: Community members should participate meaningfully rather than serve only as sources of extractable data.
• Career relevance: Roles such as data analyst, UX researcher, policy technologist, digital ethnographer, and social-impact evaluator combine disciplinary and technical capabilities.

VII. Ethical issues in AI and responsible use of emerging technologies — Rights, safeguards, and accountability

Ethics in emerging technology concerns how power, benefits, risks, and responsibilities are distributed. Responsible use requires anticipating harm, involving affected people, and maintaining accountability throughout a system’s life cycle.

A. Ethical issues in AI and responsible use of emerging technologies

The central requirement is that innovation respect autonomy, justice, privacy, safety, and human dignity.

• Privacy and informed consent: Participants should know what data are collected, for what purpose, and with whom they may be shared. Public availability does not automatically equal ethical permission for unrestricted research use.
• Bias and discrimination: Bias can enter through historical data, non-representative samples, labels, model design, or unequal deployment. Fairness checks should compare error rates across relevant groups.
• Transparency and explainability: Users need understandable information about system purpose, data sources, significant factors, and limitations, especially in decisions affecting employment, credit, education, or welfare.
• Accountability: Institutions remain responsible for automated decisions. A statement that “the algorithm decided” does not transfer legal or moral responsibility to software.
• Data governance: Data minimisation, purpose limitation, access controls, retention periods, encryption, and secure deletion reduce unnecessary exposure.
• Consent and power: Consent may be weak where participation is compulsory or where people cannot realistically refuse a public-sector or employer-controlled system.
• Misinformation and deepfakes: Synthetic text, audio, and video can manipulate public opinion; provenance checks, labelling, and media literacy are important safeguards.
• Environmental cost: Training and operating large computational systems consume electricity and hardware resources, so efficiency and lifecycle impacts should be assessed.

B. Applications and limitations

Responsible technology requires continuous evaluation before, during, and after deployment.

• Impact assessment: Identify affected groups, possible harms, benefit distribution, security threats, and routes for appeal before implementation.
• Human oversight: A trained person should review high-stakes recommendations and be able to override them with documented reasons.
• Auditability: Maintain records of datasets, model versions, prompts, decisions, incidents, and corrective actions.
• Proportionality: Use the least intrusive technology capable of meeting the legitimate objective; collecting continuous location data is excessive if periodic aggregate counts are sufficient.
• Redress: People should be informed when automated systems affect them and should have accessible procedures to challenge errors.
• Inclusive design: Test systems with varied languages, disabilities, ages, cultures, and connectivity conditions so that innovation does not deepen existing inequalities.