TL;DR
Downing Street has said Prime Minister Andy Burnham “writes most of his own speeches”, after a member of his team was pictured using ChatGPT while working on his Saturday 14 November speech to Welsh Labour’s conference. The Mail on Sunday ran the photograph. No 10 said the aide sought feedback on a speech that had already been written, and PA understands the chatbot’s suggestions were not used.
What happened
According to the Mail on Sunday, as reported by PA, the aide was on a Euston to Holyhead train and asked ChatGPT for comments on the text. The paper says the chatbot suggested ways to make a passage “a little more emotive” and said it preferred a section that moved from miners to their families and the wider community.
Downing Street’s statement read: “The PM writes most of his own speeches. An aide sought feedback on a speech draft that had already been written.”
The speech itself focused on industry and belonging. Burnham said the country “took some wrong turns in the 1980s” that broke up working communities, and described the clubs, choirs, sports teams and trade unions that Welsh workplaces once sustained.
The political reaction
Lee Anderson, who chairs Reform UK, seized on the photograph. “AI Andy’s man-of-the-people act has been exposed as a sham,” he said, joking that writing the Prime Minister’s speeches must be “thirsty work”. In the Mail’s photograph the aide appears to have a glass of Prosecco.
Looking Forward
The episode is minor in itself, and No 10’s line is that a chatbot reviewed a finished draft rather than writing it. In our view it is still a useful marker of where public norms sit. Asking AI for comments on a draft is an ordinary use of the tools, yet a political speech is treated as a test of authenticity, and opponents will read any AI involvement as a sign the words are not the speaker’s own. We expect more political offices to be asked where they draw the line on AI help with speeches and public statements, and some to set that line out in advance.