Skip to main content

beta This is the improved Camden website. Tell us what you think.

Camden's Generative AI Pilot Programme

How Camden uses data and AI

Our promise: We will only use data and artificial intelligence (AI) in ways that benefit residents, protect privacy, and are open and transparent. These promises come from Camden’s Data Charter, which was co‑designed with residents and guides everything we do. 

What is the Data Charter?

The Data Charter is a set of principles that explains how we collect, use and share data ethically. It was created with two resident panels (2021 and 2023) so it reflects what our communities expect from us. Our principles include being transparent, keeping data secure and ethical, protecting individual rights, using data for public good, and making our information clear and accessible. We also commit to publishing information such as data protection impact assessments and a Data Sharing Register.

About Camden’s Generative AI Programme

We are running a structured programme to safely explore where generative AI (GenAI) can genuinely help residents and staff. GenAI will never replace human judgement, a trained member of staff always remains “in the loop” and is responsible for all work they produce. 

Why we’re doing this

GenAI can help with repetitive writing and summarising, make information clearer, and free up our staff to spend more time on the things that matter -supporting residents and improving services. Evidence suggests that these tools can increase the quality of work produced, reduce the time taken to complete tasks, while having additional benefits to staff wellbeing and approaches to work in new innovate ways.

We intend only to adopt tools when the benefits outweigh the risks and align with our Data Charter. 

How the programme works 

  • Ongoing cycles: Running 6‑month cycles, so we can learn, improve and scale what works, if appropriate. Each cycle involves around between 100 - 200 participants, with a small number of proof‑of‑concept (PoC) projects to test deeper changes.  
  • What we measure: We track productivity improvements, quality changes, innovation against the costs of using and implementing these tools, with regular check‑ins to see what should continue or stop.   
  • Problem‑led, not tech‑led: We choose use cases that solve real service problems and assess likely value and risks before moving ahead. 

Training and access

  • Foundational training and assessment are offered to staff so they understand risks and good practice. Passing the assessment unlocks full access of approved tools.
  • Deeper training on ‘Prompt Engineering’ and ‘Critical Thinking’, and deeper dives into each tool and it’s capabilities, alongside coaching for staff who need extra support.
  • We provide clear guidance on what’s appropriate to put into AI tools. 

Tools we’re testing 

  • We explore enterprise‑grade tools such as Microsoft Copilot Chat , Copilot M365, ChatGPT, Magic Notes, Grammarly, Copilot GitHub, Adobe Firefly, Minute AI. 
  • These tools are available for assessed use cases only and all staff must follow our coporate AI policy when using them.  Camden resident data is not used for training by any company we work with as strict agreements are in place under commercial agreement

Our safeguards

  • Your data isn’t used to “train” AI. Our approved tools do not use Camden’s data to train public models; data stays within our systems. 
  • Checks before use: Every new tool or use case goes through a Data Protection Impact Assessment (DPIA) and a Camden AI Risk Assessment (AIRA), as well as information‑security checks. 

    Published AIRA’s can be found alongside DPIA Pre-Screens and full DPIA’s which outline what data we are processing via AI tools. These can be found at the following links:

    •    Pre-Screen DPIA
    •    Full DPIA

  • Human in the loop: Staff must review AI outputs for accuracy, fairness and context. We train colleagues on risks like bias and hallucinations so they can spot and correct issues.  
  • Transparency: We explain what we’re doing in line with the Data Charter and share updates with residents. 
  • On our Data Charter pages we share our principles, commitments and links to privacy information and case studies.

What we’re working on with services

Some examples of the 75 use cases currently being explored include:

  • Using AI to process user research used to provide insights into how we develop our webpages.
  • Developing an AI agent that surface important information to our call handlers to improve speed and quality of response on our phone lines. 
  • Using AI to identify and highlight themes and trends within public consultations.
  • Using AI to review communications to improve quality and consistency with citizens.

How residents benefit

  • Faster responses – GenAI helps staff handle routine queries quickly, so residents spend less time waiting for answers.
  • Clearer communication – Tools can simplify complex information into plain, accessible language for everyone.
  • More time for people – By automating repetitive tasks, staff can focus more on face-to-face support and solving individual problems.
  • Smarter insights – Staff can use GenAI to analyse information and spot trends, leading to services that better match community needs. 

What we will not do

  • We do not use your personal data to train commercial AI models. 
  • We do not let AI make decisions without a person being responsible. Staff remain accountable for decisions.
  • We do not roll out new AI tools without privacy, risk and security checks. 

Learn more and get involved