The Government Digital Service (GDS) is testing the use of AI to help government information management professionals tackle an email management challenge it says is no longer sustainable to handle entirely manually.
In a recent blog, GDS explained how Project Kestrel is exploring whether modern data science techniques can help Knowledge and Information Management (KIM) teams review large volumes of government email more quickly and consistently.
The project is led by the AI Data Readiness Incubator team within GDS and has been developed in close collaboration with the Government KIM team.
Email has accumulated across government for decades, according to the blog, encompassing business decisions, administrative threads, sensitive exchanges and personal messages.
KIM specialists can be required to work through hundreds of thousands of items, deciding what should be retained, what meets legal and compliance requirements and what may be sensitive enough to require careful handling.
GDS said that at the scale at which government operates, carrying out that work manually “simply is not sustainable”.
Creating synthetic government emails
One challenge was that Project Kestrel couldn’t safely use the data it needed for experimentation. Government emails can contain personal information, sensitive material and legally protected content, creating risks if they’re used to develop and test new tools.
If you liked this content…
Tech leaders warn against dismantling UK digital leadership as Burnham eyes Whitehall overhaul
Government launches Cloud Challenge Book to reshape public sector cloud strategy
GDS and National Archives pilot highlights importance of data maturity for AI adoption
DSIT spending plans put digital government and AI at heart of reform agenda
GOV.UK gets conversational as government launches AI chatbot
As a result, the team developed Tiger Heron, a synthetic email generator that uses large language models to produce realistic examples resembling the content government teams encounter without including real or sensitive information. This allows teams to develop and test technology without accessing genuine government inboxes.
Machine learning brings order to email
The second component of Project Kestrel is Cuckoo, a machine learning capability designed to address the scale of the email management challenge.
Within a controlled test environment, Cuckoo reads large volumes of emails and groups them into clusters according to their content. Rather than opening every message individually, KIM professionals can use an interactive visual interface to see emails grouped around similar topics, threads and types of content.
The approach is intended to help reviewers focus their attention, identify patterns and make decisions that are easier to document and defend.
Keeping humans in control
GDS said the objective isn’t to replace human judgement, but support KIM professionals by making their work “faster, more consistent and easier to explain”, while keeping people responsible for important decisions.
Phase one has also involved user research with KIM professionals, ethics input, engagement with stakeholders across government and the development of best practice materials.
GDS said the phase was designed as an experiment rather than to produce a finished product, and further testing is still required. However, it described Project Kestrel as an early example of how government can experiment with AI while addressing the practical challenge of safely working with personal and sensitive data.
