I use AI as an MP – and I’m not apologising for it
3 min read
Parliament and political parties must move more quickly to adopt AI. With the right safeguards in place, it can be a huge help for MPs.
I became an MP because I wanted to help people in my constituency and bring about positive change in the UK.
Whilst I knew there would be a significant amount of administration and constituency casework involved, nothing could have prepared me for the 500-plus emails I receive each week, ranging from template “clicktivism” campaigns run by charities to highly complex legal issues affecting constituents in Northampton South.
MPs approach their casework in varied ways. I’ve heard of some Conservative MPs who refuse to respond to emails entirely, claiming GDPR prevents it, and instead only reply by letter if a constituent writes to them by post. I’ve also heard of MPs whose staff draft every email, but who still send them out as if the MP had written them personally.
For me, I’ve chosen to employ some brilliant caseworkers who handle complex support cases as well as the broad range of campaign and policy queries we receive. But when the inbox overflows, or if I have a spare few minutes between meetings, I’m more than happy to help.
My approach to casework unexpectedly made the news this month. A former MP’s staffer sat behind me on a train and put their phone through the gap in the seats to take photos of my screen as I worked on my way back to London one Sunday afternoon. They then sent those pictures to national newspapers, hoping to create a scandal. The images showed me using the casework software that almost every MP in the UK relies on, alongside a generative AI tool called ChatGPT, to help refine my response to a constituent.
I’ve been genuinely surprised by the response to the coverage. Almost every comment has been positive. I even ended up being featured on This Morning, as Gyles Brandreth and Kay Burley discussed the story, not so much concerned about the software, but about my privacy and, crucially, the privacy of my constituents.
Now there is a wider debate on whether MPs should use ChatGPT or other generative AI tools to help draft constituent correspondence. My strong view is yes, provided safeguards are in place and a human always has the final say on what goes out the door.
In my previous career, I was fortunate enough to be part of a team rolling out AI tools, particularly in editorial writing. That experience has been invaluable in my role now. Of course, we use a closed system that ensures no data is fed back into the large language model behind ChatGPT. Of course, a human always checks, edits, and personalises the response to ensure it truly answers the question, rather than offering generic waffle. And of course, we’ve ‘trained’ our chosen AI tool with my policies, red lines, and priorities, so its first drafts reflect my principles.
My key learning from this experience is that political parties and the House of Commons need to move faster in adopting technology, and ensuring safeguards are in place to protect constituent and private data.
The government has rightly put AI at the heart of its vision, and it will transform planning and patient care in the NHS and other departments. In time, I’m sure we’ll see it transform the wider public sector for the better.
But right now, our priority must be to educate MPs and staff, safeguard data, and protect the integrity of Parliament. That way, we maximise the time available to focus on the complex work of delivering real change for the better of our country.
Mike Reader is the Labour MP for Northampton South.
