This research explores the impacts of data and AI on society and harnesses the power of new technologies to enhance research methodologies, as well as seeking to refine and improve AI capabilities. It combines technical aspects of data science and AI (statistics, machine learning and computer programming) with their applications in the social world, encompassing political, economic, social, legal, policy and philosophical questions.
There are a six key themes under which research is focussed:
Sustainability and environmental impact
The response to climate and environmental challenges is as much a social question as a physical one. How can we as individuals, organisations and societies make effective decisions to address our environmental impacts? How can we make sustainability a foundation of all that we do? Data science and AI have a potentially very significant role to play in helping us in environmental monitoring and modelling, and in helping to optimize resource use and manage energy efficiency. However, the very use of AI algorithms and the storage of enormous amounts of data has clear environmental impact. There are therefore regulatory and policy issues as well as questions around the development of ‘green AI’.
Related LSE Research
The Grantham Institute was founded in 2008 with the vision of a world in which climate change and other global environmental challenges are managed effectively to promote prosperity and well-being.
LSE Cities carries out interdisciplinary research into how cities can be designed and managed to be more open, promote social inclusion, reduce environmental impact and be governed more effectively.
Digital societies, misinformation and democracy
Data from our social existence is generated, gathered and tracked like never before. Trends towards digitisation have been sharply accelerated, changing the way we live in the twenty-first century. We must consider how social data is changing how people interact with friends, businesses, employers, politics and democracy. AI and data science have the potential to enhance democratic engagement, but there are many possible negative effects and risks to be managed—such as misinformation, erosion of trust, biases, creation of echo chambers, deepfakes, manipulation and behavioural influence, surveillance and content moderation or censorship. Balancing the benefits with the risks will be crucial to ensuring a society in which individuals can meaningfully engage and in which democratic principles are preserved.
Related LSE research
Department of Media and Communications
Department of Government
Ethics, regulation and policy
The dramatic rise of artificial intelligence, machine learning and autonomous systems raises many ethical and regulatory questions about the role of data in society. Data can contain inherent biases, and systems using that data can be flawed as a consequence. How can we ensure that decision-making using AI is ethical, transparent and explainable? With whom does accountability and responsibility lie? How can we ensure the privacy and security of personal data, and the intellectual property rights of creators of texts and other media used for training systems? As AI systems become more sophisticated, how can we ensure that human-AI interaction is constrained so as to prevent manipulation or misinformation? Significant ethical issues are raised by many of the potential applications of AI and data, such as for military uses. There are, too, important issues around the environmental impact of these technologies.
Related LSE research
The interdisciplinary Centre for Philosophy of Natural and Social Science that explores philosophical, methodological and foundational questions arising in the natural and the social sciences has published research into the ethics of artificial intelligence.
Finance, Economics and the Future of Work
The use of AI in finance presents issues around volatility and systemic risk, with a consequent need to consider regulatory reforms. Transparency, explainability and fairness will be important in developing any systems that make financial decisions (such as whether to approve loans or mortgages). The wider effects of AI on the economy -- and in particular job displacement or replacement -- raise many issues, from the need to upskill workers and develop new ways of working, to macro issues such as the possible increased concentration of wealth and the feasibility or desirability of ameliorative effects such as a universal basic income.
Related LSE research
The Systemic Risk Centre (SRC) was set up to study the risks that may trigger the next financial crisis and to develop tools to help policymakers and financial institutions become better prepared.
The Financial Markets Group Research Centre (FMG) was established at LSE in 1987. The FMG is a leading centre in Europe for policy research into financial markets.
Health, Wellbeing and social care
Immense amounts of data are available from the field of health and social care and, increasingly, machine learning and AI are playing significant roles in the analysis of such data, from identifying the need for preventative interventions, to diagnosis; and also in the administrative operations of health systems. How can we ensure that these technologies are used effectively and appropriately to support practical decisions at all scales without compromising the privacy and interests of individuals? How can we ensure trust, patient confidentiality and consent, equity of access and treatment; and where does accountability and liability rest?
Related LSE research
The Care Policy and Evaluation Centre (CPEC) is a leading international research centre carrying out world-class research in the areas of long-term care (social care), mental health, developmental disabilities, and other health issues - across the life course - to inform and influence policy, practice, and theory globally.
Methodological foundations for Social Science
The use of AI and data to support social decision-making requires careful considerations around fairness and algorithmic design, and the use of ‘black box’ methodologies rather than explainable or transparent ones. There has been much focus on AI for Science (such as applications to drug design and protein folding), but less so on AI for Social Science. Are the systems being developed optimal for social science applications, or could we work with developers to develop more useful models or algorithms? For instance, could we develop enhanced computational models of populations or societies that interact with their environments and respond to stimuli, with the purpose of using them as a tool for modelling policy interventions?
Related LSE research
The Department of Statistics focuses its research activities in the data science area in the area of developing machine learning and statistical methods, their theoretical foundations, and applications.