How AI can help journalists track MPs financial interests

Credit: Created using Magic Media, a generative AI app

Did you know that Conservative MP Jacob-Rees Mogg receives nearly £30,000 a month for presenting a show on GB News? Or that Labour MP Diane Abbott got £3,000 for a speech at a college?

All this information is publicly available in The Register of Members’ Financial Interests, an online website updated fortnightly that tracks money that UK MPs receive on top of their salaries.

Its main purpose is "to provide information about any financial interest which a Member has, or any benefit which he or she receives, which others might reasonably consider to influence his or her actions or words as a Member of Parliament.” Or, put simply, it looks at money that may influence British politics.

The caveat is that the register is incredibly hard to navigate, says Clare Spencer, research and development producer, BBC News, UK, at the JournalismAI Festival today. So she joined forces with Juan Ginzo, principal data scientist, at The Times to create a tracker that uses generative AI to help journalists use this resource more easily.

• Read more: I am a public service journalist: what can AI do for me?

MP Interests Tracker is a research tool that helps journalists analyse and compare data that is often inconsistent. For instance, MPs can write company names in different formats, or phrase their interests in various ways. This makes it harder to make connections between the politicians and the companies and understand how the money flows.

The tool extracts the raw register entries and breaks them down to clearer data, using Open AI GPT-3.5 as a backend. It then creates a structured version of the data that journalists can analyse to find those high-ticket items or tie MPs to specific companies and industries.

Ginzo adds that MP Interests Tracker works well for surfacing ideas for stories and finding those needles in the haystack. However, it is not 100 per cent reliable and it may miss some important data or get false negatives.

However, like with anything produced by generative AI, all results should be checked against the source for accuracy.

According to the creators, the tracker is easy to prototype. They are currently testing it and once it is in a good enough shape, they want to turn it into a product.

Free daily newsletter

If you like our news and feature articles, you can sign up to receive our free daily (Mon-Fri) email newsletter (mobile friendly).