Scottish AI Register

Search and filter the register

Stage

Showing 25 to 30 of 34 systems

AI systems

  • Overview — Codini is an AI-powered coding platform (developed by EvooSurvey) used to analyse open-ended survey responses through thematic coding and sentiment analysis. It applies natural-language processing to identify themes and sentiments, supporting efficient and consistent qualitative analysis. How it is used — Progressive Partnership uses Codini to analyse anonymised open-ended survey responses collected from the Scottish public as part of a Scottish Government-commissioned study on circular-economy behaviours. Use is limited to Scotland and focused on public attitudes and practices related to the circular economy. It is used only once survey responses have been anonymised and prepared for analysis. Scope and limitations — Codini supports thematic coding and sentiment analysis to help identify patterns and themes in qualitative data. It does not perform automated decision-making, profiling or predictive modelling, does not generate insights independently, and does not interact directly with data subjects. Its role is limited to assisting researchers in organising and interpreting qualitative data. Human oversight — Researchers at Progressive Partnership review and refine the AI-generated coding to ensure accuracy, and data is deleted on completion of the project.

  • Overview — This proof of concept examines whether AI can assist with the redaction of sensitive information within Scotland's Children's Hearings System, the statutory system of care and justice that makes decisions about compulsory measures of supervision for children and young people in need of care, support and protection. It forms part of SCRA's wider AI programme, which identifies and tests potential applications of AI in the Children's Hearings System, based on engagement with staff about how AI might address common challenges and free up staff to better support children and families. How it is used — SCRA safeguards children partly through non-disclosure provisions, which allow information that would otherwise be shared to be withheld from an individual, for example withholding address information where a child's whereabouts must not be disclosed to a parent, by redacting it from materials issued for Hearings. In this proof of concept, the AI scans the text of multiple documents to recognise patterns of words and phrases that should not be disclosed and highlights them as suggestions to the staff carrying out redaction. It draws on the non-disclosure information held in SCRA's case-management system, such as a child's, parent's or carer's name or address, and places such as their school or medical centre, and learns to identify patterns indicating the whereabouts of a child, applying this across other reports and documents. Human oversight — The redaction itself is not automatic. Staff check the AI's suggestions and remain responsible for reading through reports to identify any information not previously considered for redaction, such as the name of a sports or youth group a young person attends. Overall accountability for redaction remains with human decision-makers. Assessment and impact — The tool will be tested on 50-100 cases over three months to establish how long AI-based redaction takes, how accurately it identifies the required information, the number and causes of any data breaches, and staff confidence in using it. As with the related proof of concept, a wider research project is gathering the views of children, families, carers, social workers, safeguarders, advocates, solicitors, SCRA staff and panel members on the ethics, rights and legalities of using AI, and equality and human rights, children's rights and data-protection impact assessments have been undertaken.

  • Overview — This proof of concept examines whether an AI tool can support the preparation of accurate witness statements within Scotland's Children's Hearings System, the statutory system of care and justice that makes decisions about compulsory measures of supervision for children and young people in need of care, support and protection. It forms part of SCRA's wider AI programme, which identifies and tests potential applications of AI in the Children's Hearings System, based on engagement with staff about how AI might address common challenges and free up staff to better support children and families. How it is used — When a child or young person is referred to SCRA, a Children's Reporter assesses whether there is evidence that compulsory intervention is needed and, if so, arranges a Children's Hearing; where the child or family does not accept the reasons, SCRA may go to court, and the Reporter asks individuals concerned about the child to give evidence. In this proof of concept, Children's Reporters use a web-based transcription tool to record and transcribe interviews with prospective witnesses, and the transcripts are turned into witness statements in line with SCRA's practice standards. Human oversight — Although the initial witness statement is generated automatically, both the Children's Reporter and the witness are responsible for verifying its accuracy: the Reporter checks and corrects the transcript and edits the statement to ensure it is accurate and relevant, and the witness is given a copy to confirm it accurately reflects their evidence. Assessment and impact — The tool has been piloted by at least 14 councils in England and Wales, where evaluations in social-work services indicated high accuracy and a significant reduction in administrative burden. SCRA is further assessing its accuracy in the Scottish legal context through focus groups with Children's Reporters, the logging of errors fed back to managers and developers, and a research-team assessment of transcription accuracy across different accents, ages, genders and ethnicities, compared with other transcription services and human transcription. In parallel, a wider research project is gathering the views of children, families, carers, social workers, safeguarders, advocates, solicitors, SCRA staff and panel members on the ethics, rights and legalities of using AI, and equality and human rights, children's rights and data-protection impact assessments are being undertaken.

  • Overview — Cursor is an AI-enabled integrated development environment (IDE), built as a fork of Visual Studio Code, that analyses full project context to offer code completions, multi-line refactorings, chat-style code queries and edits within the developer's editor. How it is used — It assists developers by providing context-aware code generation, bug detection and refactoring. It operates under three privacy modes (No-Storage, Privacy Mode and Standard) and holds SOC 2 Type II certification; privacy and usage settings can be managed per team, with Business and Enterprise plans enforcing full privacy mode by default. Scope and limitations — Cursor is designed to assist developers with context-aware code generation, bug detection and refactoring. It is not designed to store or train on user codebases when Privacy Mode or No-Storage Mode is enabled. Human oversight — Developers remain fully responsible for reviewing and accepting AI-generated suggestions; Cursor does not autonomously commit code, and human review remains mandatory. Concerns about suggestion accuracy or privacy incidents are handled through existing code-review and IT escalation processes, and issues can be escalated to Cursor support if required.

  • Overview — Amazon Q Developer is a generative-AI assistant (built on Amazon Bedrock foundation models and augmented with AWS-specific content) that supports software development and operations. It provides agentic coding, code generation, inline completions, unit-testing suggestions, vulnerability scanning, documentation generation and infrastructure analysis. How it is used — It works within IDEs (such as VS Code, JetBrains, Visual Studio and Eclipse), the AWS Management Console, the AWS Console Mobile app, the command line, and chat apps such as Teams and Slack. It is used to assist with AWS application development, operations, resource troubleshooting, code refactoring, documentation creation and cloud cost analysis, with access controlled through AWS IAM governance and managed service-linked roles. Scope and limitations — It is not intended for autonomous decision-making; developers retain responsibility for final code and actions. Human oversight — Developers remain responsible for reviewing and approving all code generated by Amazon Q Developer, which does not autonomously make production commits, so human supervision is required and the system does not override user decisions. Concerns about accuracy, security suggestions or behaviour are escalated through existing developer-review and AWS support channels, with governance managed via IAM identity and internal escalation protocols.

  • Overview — MaxQDA with AI assist is qualitative-analysis software used to code and analyse textual research data. The Social Research Profession is trialling its AI-assisted features to evaluate whether they can speed up qualitative analysis and improve its quality. How it is used — Within the profession's ethical guidelines, the tool is being piloted across a range of analytical scenarios using qualitative (free-text) social-research data such as consultation responses and transcriptions of interviews and focus groups. Researchers use it to code data and query those codes in thematic analysis; the AI features can generate summaries of codes and coded text and support querying the documents to understand them better. Scope and limitations — The trial is restricted to MaxQDA with the AI assist add-on; other add-ons (Transcription, Team Cloud and Tailwind) are not included. Social-research policy is that the tool is used only as an adjunct, secondary to the primary analysis, which is conducted by qualified researchers.