UK Data Concerns and AI Applications for Climate, 2024

Data were collected through a survey instrument administered by a market research company (Qualtrics) in April 2024 to a nationally representative sample (age and sex) of adults in the UK. The UK was chosen as the reference country due to the wide availability of climate-relevant applications and the policy imperative to reduce per capita emissions in line with legally binding domestic climate targets. Participants received monetary compensation and the final sample size is 2,078.
The survey instrument comprises of six blocks of questions followed by a vignette-based experimental survey block. Questions used either single or multi-item scales based on precedents in the literature (with slight modifications to fit the current research context) and newly developed items. Many questions consisted of statements with level of agreement captured by a 5-point Likert scale. The median duration of the survey was 11.9 minutes.
Initial survey blocks captured demographic and contextual characteristics to inform inclusion criteria (above the age of 18) and quota fulfilment (nationally representative sample on age and gender), followed by self-reported use of a diverse set of 15 data-driven climate-relevant applications largely enabled by AI (e.g., learning algorithms, cloud-based services, or other forms of data-dependent service-provision) in four domains of daily life: retail, mobility, food and home energy.
These applications are referred to as ‘climate-relevant’ to denote applications whose design and use have implications for energy demand and emissions, recognising that their ultimate climate impact depends on user behaviour, system interactions, and potential rebound effects. Logic branching allowed follow-ups regarding usage frequency to be customised according to current, past or non-use. For example, a participant who indicated current use of an application was then asked, ‘how often do you do [activity]?’.
Subsequent blocks assessed usage propensity for four case study applications (one per domain), general attitudes towards technology, data privacy concerns, data protection behaviours and perceptions of and attitudes to AI. The latter was taken directly from the Office for National Statistics (ONS) survey on AI and included a measure of perceived AI risk which did not specify a particular type of risk. This block was positioned after questions on data privacy concerns and protection behaviours to provide contextual grounding in data-related issues. Several items on AI information sources were adapted from,  and additional questions included, for example to assess AI familiarity e.g., usage of GenAI.
To examine the potential effects of the recent surge in AI awareness following the release of ChatGPT 3.5, we also included retrospective items measuring changes in attitudes and behaviours over the previous year of 2023 (the first full year of the GenAI hype). To capture change in application usage, we asked ‘Compared to a year ago, how would you describe the frequency you do the following activities?’.
Following the main survey, we implemented a vignette experiment to test how awareness of AI data practices affects perceptions of climate-relevant apps that use AI and their usage propensity. The vignettes represented stylised but plausible versions of existing or emerging applications rather than descriptions of specific commercial products. They were designed to capture perceptions of AI-related data practices across realistic, recognisable domains (e.g., smart thermostats, bike-sharing, food-waste apps) while maintaining experimental control. To ensure realism, we mapped each vignette to a representative market leader and reviewed their publicly available documentation on ownership, business model, data collection, and AI use.
Each respondent was randomly assigned to one of the eight vignettes, describing the use of one of four climate-relevant applications, with either an increased or reduced emphasis on personal data being collected by AI. The vignettes respected the principles of realism, clarity, simplicity and internal consistency.

The Data Collection is available for download to users registered with the UK Data Service. All requests are subject to the permission of the data owner or his/her nominee. Please email the contact person for this data collection to request permission to access the data, explaining your reason for wanting access to the data, then contact our Access Helpdesk.