FOI 1024-2024/25 AI Strategy

AI Strategy and Planning

Q1. To what extent does your department have a clearly defined and documented AI strategy? Please select one of the following options that applies. 

• No AI strategy in place.
• Initial discussions or ideas about AI, but no concrete plan.
• A basic AI strategy is in development or has been drafted.
• A well-defined AI strategy is in place, but implementation is in the early stages.
• The AI strategy is being actively implemented with measurable progress.
• A fully defined and implemented AI strategy is driving significant impact.
  
  

 Q2. To what extent does your AI strategy or goals align with the AI opportunities action plan? Please select one of the following options that applies. 

• Unaware of the AI opportunities action plan or no attempt to align with it.
• General awareness of the AI opportunities action plan, but no specific efforts to align.
• Some aspects of the AI strategy or goals are aligned with the AI opportunities action plan.
• Conscious effort made to align the AI strategy with the key principles and directions of the AI opportunities action plan.
• Strong alignment between the AI strategy or goals and the AI opportunities action plan, with clear connections and shared objectives.
• AI strategy or goals are fully integrated with the AI opportunities action plan, contributing directly to its successful implementation.

Q3. What is the likelihood that your department's AI strategy will be fully implemented within the next two years? Please select one of the following options that applies. 

• Significant obstacles and uncertainties make full implementation unlikely in the next two years.
• Implementation is in progress but faces challenges that may delay completion.
• Moderate confidence in achieving full implementation within the timeframe.
• High likelihood of full implementation, with most key elements in place.
• Almost certain to achieve full implementation within the next two years.

Q4. How confident is your department that current AI governance structures are adequate for managing AI risks? Please select one of the following options that applies. 

• No AI governance structures in place or serious concerns about their adequacy.
• Limited AI governance in place, with significant gaps and areas for improvement.
• Moderate confidence in current AI governance, but some areas of uncertainty remain.
• Generally confident in the adequacy of AI governance structures to manage most risks.
• Strong confidence in AI governance, with robust processes and oversight in place.
• Extremely confident that AI governance structures can effectively manage all foreseeable AI risks. 

Q5. What percentage of your department's total budget is allocated to AI-related projects this fiscal year? How much is this? 

Q6. How does your department or agency assess the economic impact or cost-effectiveness of AI technologies? Please select one of the following options that applies and provide any relevant data or reports on AI’s economic benefits. 

• No formal assessment: No systematic approach to evaluating the economic impact of AI.
• Qualitative assessment: Relies on anecdotal evidence, case studies, or expert opinions to assess the impact of AI.
• Quantitative assessment: Uses metrics, data analysis, and modelling to measure the economic benefits and costs of AI.
• Comprehensive evaluation: Includes both qualitative and quantitative assessments, considering a wide range of economic factors and potential impacts.

 AI Deployment and Use Cases

Q7. What is your department's current stage of AI deployment? Please select one of the following options that applies. 

• No AI use cases.
• Exploring opportunities
• Planning for at least one AI use case.
• Piloting at least one AI use case.
• At least one AI use case is fully deployed.
• Multiple AI use cases fully deployed.

Q8. For your most significant AI use case, how confident are you that the expected benefits will be realised? Please select one of the following options that applies. 

• No clear expectations of benefits or serious doubts about achieving them.
• Some anticipated benefits, but significant uncertainty about their realisation.
• Moderate confidence in achieving some of the expected benefits.
• Confident that most of the expected benefits will be realised.
• Strong confidence that the AI use case will deliver significant benefits.
• Extremely confident that the AI use case will fully achieve or exceed its intended benefits.

 Risk Management and Ethics

Q9. How thoroughly has your department assessed the risks associated with AI deployment? Please select one of the following options that applies. 

• No formal risk assessment conducted for AI deployment.
• Initial discussions about AI risks, but no systematic assessment.
• Some areas of AI risk have been assessed, but gaps remain.
• A moderate risk assessment has been conducted, covering key areas.
• A comprehensive risk assessment is in place, addressing a wide range of potential risks.
• A thorough and ongoing risk assessment process is embedded in AI development and deployment. 

Q10. How confident are you that your department is compliant with relevant data protection laws when using AI? Please select one of the following options that applies.

• Significant concerns about compliance with data protection laws.
• Limited understanding of data protection requirements for AI.
• Some measures in place to ensure compliance, but gaps may exist.
• Reasonably confident in compliance with data protection laws.
• Strong confidence in data protection compliance, with robust processes in place.
• Fully confident in compliance with all relevant data protection laws and regulations. 

Q11. How well does your department ensure transparency in the use of AI in public services? Please select one of the following options that applies. 

• No efforts to ensure transparency in AI use.
• Limited information provided to the public about AI systems.
• Some transparency measures in place, but room for improvement.
• Reasonable transparency in AI use, with public information available.
• Strong commitment to transparency, with clear communication about AI systems.
• Full transparency in AI practices, including open data and explainable AI.

Data Residency and Sovereign Infrastructure

Q12. How does your department or public body ensure that its data storage complies with UK sovereignty and data residency requirements? Please outline any policies or regulations guiding data storage within UK borders.

Q13. Can you provide the proportion of data stored within UK-based infrastructure versus foreign cloud services (e.g., US-based providers) within your department/agency?

Q14. What measures are in place to ensure that your data storage solutions meet national security standards? Please provide details of relevant frameworks or protocols currently implemented.

Energy Consumption and Efficiency of AI

Q15. What is the estimated carbon footprint of your department or agency’s AI operations (e.g., training, deployment) over the past year? Please include energy consumption figures if available. 

Q16. How do you track the energy usage of AI workloads, and are there any reporting mechanisms in place for sustainability goals? Please provide relevant metrics or guidelines.

Data and Infrastructure

Q17. How would you rate the quality of data used for your AI systems? Please select one of the following options that applies. 

• Data quality is poor and unreliable.
• Data quality is a concern, with significant issues.
• Data quality is adequate but could be improved.
• Good data quality with regular validation processes.
• High data quality with robust quality assurance measures.
• Excellent data quality with continuous validation and improvement.

Q18. What is the level of confidence in your department's digital infrastructure's ability to support current and future AI initiatives? Please select one of the following options that applies. 

• Serious concerns about the infrastructure's ability to support AI.
• Infrastructure limitations may hinder AI initiatives.
• Moderate confidence in the infrastructure's capacity for AI.
• Confident that the infrastructure can adequately support current and near-term AI needs.
• Strong confidence in the infrastructure's ability to support ambitious AI projects.
• Extremely confident that the infrastructure is future-proof and ready for any AI challenge. 

Q19. Does your department or agency currently use a hybrid infrastructure (i.e., a mix of on-premises and cloud services) for its digital operations, including AI applications? Please specify the proportion of services hosted on each type of infrastructure. 

Q20. What factors influence your department’s decision-making process when choosing between public cloud, private cloud, or on-premises infrastructure for AI-related projects or services? 

Q21. What plans or initiatives are in place to ensure your department’s digital infrastructure is scalable enough to meet future demands for AI-powered services or technologies over the next 5–10 years?

Q22. How does your department or agency ensure the quality of data used in AI-driven projects or services? Please provide any metrics, guidelines, or standards employed for data validation. 

Skills and Training

Q23. How effectively is your department upskilling employees to use AI? Please select one of the following options that applies. 

• No upskilling initiatives in place for AI.
• Limited training available, with little focus on AI skills development.
• Some training programs offered, but not comprehensive.
• Effective upskilling programs are in place, providing employees with essential AI skills.
• A strong focus on AI skills development, with comprehensive training and support.
• A culture of continuous learning and development, ensuring employees are equipped for the AI-powered 

Q24. Will you be relying on in-house skills for AI strategy development and implementation or seeking third-party services/consultancy?

Future Ambitions

Q25. How confident are you that your department will reach its AI ambitions? Please select one of the following options that applies. 

• Significant doubts about achieving AI ambitions.
• Low confidence in reaching AI goals.
• Moderate confidence in achieving some AI ambitions.
• Confident in achieving most AI goals.
• Strong confidence in reaching AI ambitions, with clear plans in place.
• Extremely confident in achieving and exceeding AI ambitions.

Extent and Result of Searches to Locate Information
To locate the information relevant to your request searches were conducted within North Yorkshire Police.

I can confirm that some of the information you have requested is held by North Yorkshire Police. 

Decision
Please note that FOI relates to recorded information held by a public authority and as such, it does not extend to providing explanations unless the answers are already held in a recorded form. This was explained at the following Tribunal:

"Information is defined in section 84 of the Act as 'information recorded in any form'. The Act therefore only extends to requests for recorded information. It does not require public authorities to answer questions generally; only if they already hold the answers in recorded form. The Act does not extend to requests for information about policies or their implementation, or the merits or demerits of any proposal or action - unless, of course, the answer to any such request is already held in recorded form." (Day vs ICO & DWP – EA/2006/0069 Final Decision). 

You may wish to refer to the ICO guidance on FOI - How to access information from a public authority | ICO. 

Some of the questions you have asked require opinion or judgement and are not based on recorded information held by North Yorkshire Police. However, on this occasion, as a gesture of goodwill and outside of the Act, North Yorkshire Police are able to provide answers as follows:

Q1. No AI strategy currently in place.

Q2. General awareness of the AI opportunities action plan, but no specific efforts to align.

Q3. Not applicable as per our response to Q1.

Q4. Moderate confidence in current AI governance, but some areas of uncertainty remain.

Q5. 0%.

Q6. Quantitative assessment: Uses metrics, data analysis, and modelling to measure the economic benefits and costs of AI.

Q7. Planning for at least one AI use case.

Q8. Confident that most of the expected benefits will be realised.

Q9. A comprehensive risk assessment is in place, addressing a wide range of potential risks.

Q10. Strong confidence in data protection compliance, with robust processes in place.

Q11. Reasonable transparency in AI use, with public information available.

Q12. Cloud solutions are investigated and DPIA and data processing contracts are completed along with NCSC cloud security principle  assessments.

Q13. 100%.

Q14. All cloud solutions are checked against NCSC Cloud Security Principles.

Q15. No information held.

Q16. No information held.

Q17. High data quality with robust quality assurance measures.

Q18. Confident that the infrastructure can adequately support current and near-term AI needs.

Q19. 10% Cloud, 90% On Premises.

Q20. Type of data being stored, criticality of the system, cost, security.

Q21. We have a 10 year capital plan and revenue plan for cloud charges. This is driven on data consumption rates, project pipelines and planned system updates and procurement exercises.

Q22. Power bi dashboards are used to check and validate system data against. Structured dip sampling exercises are undertaken regularly.

Q23. No upskilling initiatives in place for AI.

Q24. We will use online training for general awareness and vendor provided training for specific use cases.

Q25. Moderate confidence in achieving some AI ambitions.

Please note that systems used for recording information are not generic, nor are the procedures used locally in capturing the data.  It should be noted therefore that this force’s response to your questions should not be used for comparison purposes with any other responses you may receive.