Artificial intelligence
Research Topic
Language: English
This is a research topic created to provide authors with a place to attach new problem publications.
Research problems linked to this topic
- How can we better understand novel uses or applications of AI in the geospatial ecosystem, such as in the analysis of Earth Observation and Population Movement data, 3D visualisation, and climate modelling?
- What new and emerging technologies (including cloud, Artificial Intelligence, Machine Learning, and Augmented Reality/Virtual Reality) will impact geospatial skills and innovation, and access to geospatial data in the future, and how could the UK leverage these technologies?
- How do we design public-serving autonomous systems to be fair and inclusive?
- How can AI be used to identify harmful content?
- In which channels is harmful Gen-AI content most prevalent, how does it spread, and how can friction be introduced to these channels?
- What risk is there that generative AI evolves such that the content it generates can avoid detection faster than tools can be developed to detect it? How can international and industry collaboration limit this risk?
- How will the use of generative AI to create ‘deepfakes’ that manipulate people’s likeness (face, body, voice) evolve? What is the psychological impact of being deepfaked, and what harmful uses (e.g. intimate image abuse, fraud, reputational damage) will develop and increase?
- How will AI affect existing kinds of harmful online content (e.g. online abuse, scams) and what new kinds of online harmful content might it give rise to?
- Which harmful online uses of AI are likely to increase? What could be the impact of AI-generated content on attitudes, beliefs, behaviours or psychological wellbeing?
- AI will democratise access to capabilities that used to be expensive or hard to access, and create new capabilities that didn’t previously exist. As barriers (e.g. technical skills, access to specialist equipment) are reduced, AI use will increase. What is the prevalence of AI generated content online?
- One expected impact of AI will be on trust in information. How might AI reduce public trust in information available online? Do UK citizens trust AI-generated online content?
- In what ways will AI exacerbate the spread of mis/disinformation and is mis/disinformation spread by AI likely to be more effective in influencing UK audiences?
- How can policing advance its interconnectivity both within policing (e.g., AI supported call and response routing), on multi-modal devices, and across organisations?
- What are the value and limits of emerging AI technologies such as ChatGPT in policing?
- How can policing exploit advanced data analytics to create new support processes in areas such as Out of Court Disclosures, personnel vetting, and document summarisation?
- How can the police service further develop capabilities in automatic redaction and selective extraction from phones?
- What tools are available to support the police service with compliance to analytic or data governance standards?
- What legal and ethical challenges does the use of autonomous systems in policing face as robots become more able to operate independently?
- How can policing use advances in robotics to reduce or remove the need for police officers to enter hazardous environments e.g., water, fire, electrical, natural disaster, CBRN (chemical, biological, radiological and nuclear)? Further, how can a seamless and secure operation be enabled in such environments?
- What technologies can mitigate work-related trauma experienced by police staff? For example, using computer vision technologies to reduce manual assessments of child exploitation images.
- What technologies can be used to prevent crimes online, including the online mobilisation towards violence and terrorism?
- What are the opportunities of emerging technologies (quantum and AI) to revolutionise our ability to map underground assets?
- Which areas of research on the uses of next generation networks may need policy interventions (e.g. Internet of Things and Artificial Intelligence)? This may include driving strong take-up of fibre and 5G/6G, encouraging the adoption of the products and services, and increasing Willingness to Pay and supporting industry to make the necessary investments?
- Evaluate the technologies that will drive smart networks: evidencing the utilisation of the UK’s lead in AI and Edge technology to develop self-organising, secure and highly optimised network software.
- Evaluate the technologies that will drive clean networks/power efficiency: harnessing semiconductors and AI to drive more efficient telecom radios and network optimisation.
- How might AI contribute to future spectrum regulation/management?
- Detailed data on companies that specialise in the provision of AI services in life sciences.
- Does the cyber security of AI models need to follow any novel principles that aren’t set out under existing policy and technology security principles? If so, what are these measures and how do the differ from what exists? How do the vulnerabilities/risk of AI model security differ from existing cyber threats?
- How might automation, machine learning/AI change the way in which cybersecurity services are currently delivered? Do these changes lead to a reduction or even an increase in demand for cyber security skills, products and services
- How might Artificial Intelligence tools such as language models be useful in:i) processing delivery data ii) increasing understanding of geospatial areas and associated risks of delivery?
- Detailed company data on both current expenditure and investment in AI.
- Which risks from AI are the most urgent to mitigate?
- How can we apply the defence in depth approach to preparing for risks from AI?
- What human systems are resilient to impacts from AI and which are less so?
- How can we attribute the role that AI had in causing a particular harm, rather than something else?
- What are the possible scenarios for various AI risks 1, 3, 5, 10 and more years from now?
- Which data sources or key indicators should we be watching that may indicate major changes in the risk assessment, or new risks hitherto unidentified arising?
- What risk assessment methods are best suited to risks from AI?
- How can we reduce bias when using AI?
- How can we ensure use of AI is ethical?
- What is needed to enable the public sector to adopt AI?
- Will adoption of AI in key delivery departments contribute to more efficient and effective public services?
- Will investment in AI lead to reduced public sector costs in the long run?
- How can we ensure AI increases public sector productivity?
- Agile and responsive skills system: What changes in the skills system are required to meet government ambitions to support and grow critical technologies set out in the Science and Technology Framework? How can we future proof the workforce by giving them the right skills to fully embrace AI and its potential?
- Compute: What are the latent needs of the UK’s AI ecosystem for compute resource?
- Skills: How do we make sure the right skills are available to maintain a world-leading position in AI? And how do we ensure the labour force has the right skills to support individual opportunities?
- "R&D : In which areas of AI R&D is the UK strongest? What are the most significant AI R&D opportunities for the UK? Which government interventions are most effective for boosting UK AI R&D (relative to such goals as economic growth, productivity and security)?"
- How will AI impact societal outcomes, especially regarding inequality, health and the environment?
- How can we ensure public attitudes to AI are positive, and maximise trust in safe AI?
- To what extent are the potential risks posed by highly capable AI systems a barrier to economy-wide adoption, and how could progress in AI safety overcome these barriers? How can we ensure that the UK population has the right AI skills for life and work?
- How can we better understand the barriers to AI adoption?
- How will AI impact competition and innovation?
- What will the future of AI look like within the UK, and how can we monitor our progress towards the many possible scenarios?
- How should the UK position itself in terms of the global AI market? What sort of AI businesses should we particularly be looking to attract?
- Quality or quantity: Does AI enable the delivery of better outputs and increased quality from firms and businesses?
- Micro productivity: To what extent does AI impact firm level productivity?
- Macro productivity: To what extent does AI impact national productivity?
- Wages: What would be the impact of AI deployment at work on wages and costs for employees and employers?
- Productivity: What are the possible direct and indirect productivity impacts of AI?
- Indicators: What are the rapid indicators of AI impact on the labour market?
- Sectors: Which sectors are more likely to benefit from AI and which are more at risk from the downsides?
- Occupations: What are the characteristics of occupations that put them more at risk of replacement/change or give them a comparative advantage? Over what time frame will they be impacted?
- What innovative approaches to data in education could increase staff capacity and reduce workload?
- What models of management and professional development of teaching and nonteaching roles support efficient and safe use of data and technology including AI?
- What approaches or innovation are needed to support the efficient handling of data within education settings?
- What are the most robust methodologies for assessing the effectiveness of technology used for education?
- How can we adapt research methodologies to robustly measure the impact of technology in education, given its fast-moving nature?
- How can the impact of digital technology be robustly measured, and implemented in a way that supports teachers and learners?
- How has the increased accessibility of generative AI influenced HE and FE providers and students?
- How can AI and other emerging technologies be implemented in education settings so that they do not widen existing inequalities or create new inequalities?
- What are the most effective approaches to upskilling the education workforce to use AI well? What impact could this have on productivity?
- What are the best ways to ensure that AI is used safely, ethically, and in ways that protect the data and interests of children, young people, teachers, and schools and colleges? What forms of regulation and enforcement may be appropriate?
- How do AI and other digital technologies support existing ways of working in schools and colleges? What are the main opportunities for the future?
- In what ways can AI and other digital technology reduce teacher workload and improve student outcomes? How can AI and digital technology impact on productivity?
- What are the potential long-term opportunities and challenges of AI use in education at all stages?
- What are the potential impacts of AI, and how can new technologies be used safely and effectively within education?
- What is the stock of skills in the economy, where are there mismatches between need and availability? Where will the greatest skills needs be across the medium (5-10 years) and long-term (10 to 20+ years)? What impacts might we expect AI to have on future skills needs?
- What scope does data analytics and AI have to tailor services to claimants’ needs? What are the benefits and risks of digital services? What is the effectiveness of digital transformation in driving efficiency and improving satisfaction?
- In security applications, how can we rely on AI to show us all the possible threats (not seeing what we are not shown)?
- How can we ensure Situational Awareness for different human in the loop actors in autonomous aviation?
- What training needs to be delivered to interact and challenge meaningfully AI algorithms? How can we prevent skills deterioration?
- What are the public needs for explanability in AI?
- How can new approaches and technologies be applied to deter, detect, and disrupt the misuse of drones?
- What are the key factors regarding public trust on autonomous systems?
- What are the human-machine interface (HMI) requirements for AI applications such as machine vision? How can we limit overreliance?
- What frameworks can we use to ensure proportional, safe, and trusted applications of AI in the transport sector?
- How should autonomous aviation systems communicate with ground control and each other?
- What is substantial change to trigger re-evaluation of AI?
- What is the role of remote operation (assistance, decision making & control)? What are the skills and requirements for such operation for autonomous systems?
- Can we have AI learn “on the job” safely? For example, in machine vision applications.
- How can we use digital twins to increase resilience, responsiveness, and integration of our network (cross modally)?
- How can artificial intelligence, machine learning, simulation, agent based modelling and other leading data science techniques contribute to better understanding of trade and investment patterns?
- What are the risks and opportunities to creative business growth posed by new technology, including automation such as the use of AI?
- How can robotics expedite policing activities or complement the existing provision, for example, to effectively support forensic specialists identify, record, or assess marks across a crime scene?
- What resources are required to ensure the safe and efficient handling of data in education settings?
- Changing world: How have evolutions in our statistical system (such as the greater focus on administrative sources for statistics) influenced how statistics are produced, used, and valued? How may advances in wider society (such as the increasing sophistication of large language models) influence how statistics are produced, used, and valued?
- What is the expected growth of the UK's maritime autonomy and remote operations sector and what impact will this technical change have on the workforce?
- How do we account for AI-first assumptions/errors (that humans would not make)?