Why we use AI
I run a small office. A few staff cover policy, legislation, media, correspondence, events and electorate work for the whole of the ACT. The major parties have party research units, policy staff shared across a caucus, media teams, and colleagues who can each take a portfolio. As an Independent, my small team and I work incredibly hard to deliver for the ACT.
In a typical week my office receives hundreds of emails, alongside phone calls, meeting requests, invitations and casework.
Used carefully, AI lets us do more with what we have: read more of what is put in front of us, draft more, and respond to more of the people who get in touch.
AI is a tool my office uses. It is not a substitute for the judgement that sits behind anything I say or do.
How we use AI
This page sets out how my office uses generative AI. It covers all of our work: parliamentary, electorate and campaign.
My office uses generative AI tools to assist with research, drafting, correspondence, editing and reviewing. In practice, that means work such as summarising a long report or submission, preparing a first draft of routine correspondence, checking a draft for errors and inconsistencies, or organising research material. It also includes polishing language and restructuring writing, designing and delivering websites and developing notes.
These tools are used through office accounts.
What we don't use it for
We do not use AI to:
-
influence my decisions, including how I vote
-
make final decisions
-
write personal statements attributed to me
-
write opinion pieces
Your information
My office does not put personal information into generative AI tools. That includes names, casework details and personal identifying information. We practise strict data minimisation, and we do not share your personal information with any AI tool.
A person is accountable
Every output is reviewed and approved by a person before it is used or published. Statistics, legal claims and citations are checked by a person against the original source. AI tools get these things wrong, and the responsibility for catching that sits with my office.
We retain full accountability for the accuracy and appropriateness of everything we produce, whether or not an AI tool was involved in preparing it.
Telling you when we have used it
Where AI has substantially assisted in producing published content, we say so. By substantial, we mean a tool has drafted or shaped a significant part of the material. Routine assistance: checking spelling and grammar, reformatting, or summarising a document for internal use, may not be disclosed.
We do not publish images, audio or video with images of people generated by AI without clear disclosure on the material itself.
When we get it wrong
If we publish something an AI tool has got wrong and we have not caught it, we will correct it.
Review and contact
This policy is reviewed periodically. If you have questions about it, or about a particular piece of content, contact my office at [email protected].