There was a certain inevitability about Chris Cook becoming head of artificial intelligence (AI) initiatives at the Financial Times.
Apart from more than 17 years' experience as a journalist on the FT, the BBC's Newsnight programme and Tortoise Media, he was always very interested in data, computers and the big social policy issues that AI can tackle so well.
As far back as 2010, when he became the FT's education correspondent, Cook realised that it would need more specialist computing power to cover such a scattered beat with 25,000 schools, 150 universities and thousands of colleges of every kind.
He asked for a specialist computer. The FT's IT department could not help and he remembers sitting on the floor of the public policy room, building his own specialist computer from scratch, including the very big memory and fast processors needed for what he planned to do.
"I was really the only reporter approaching the story (education) from a data point of view," says Cook who says he could now do what he did then with the help of a standard laptop and the use of the cloud.
It was all good experience for when he returned to the FT in 2021 after five years as Newsnight's public policy editor and then a spell at Tortoise.
"When I came back, I thought we should do more high tech investigations rather than a traditional approach to investigations," says Cook who notes the FT has had a chequered approach to investigative reporting in the past with teams set up and closed down several times over the years.
Using tech in the newsroom
The FT editor, Roula Khalaf, who took over six years ago, was very keen on taking advantage of technological change to spur investigations, including the use of satellite photography in a new visual investigations unit.
"I was part of the first thing we did, a story on the Russian economy and whether the economy was over—heating," says Cook who leads a group of reporters, engineers and data scientists.
Against the background of the continuing war in Ukraine, the FT team collected more than 600,000 Russian job ads a month — in the end, a total of 10 million were collected and categorised — and they were able to track wage rates against the economy.
Apart from proving that such a task was possible, the study demonstrated that the Russian economy was indeed overheating.
"What we can see now is that the enormous surge in wages (because of a shortage of manpower) is no longer continuing and pay rates are slowing down," Cook explains.
Although many Russians feel richer for now, higher inflation and a cost—of—living squeeze could lead to greater social instability.
"The pressure is on the Kremlin right now and it's a big macro story that would not have been possible to tell any other way," Cook believes.
It's also a very FT sort of a story.
Cook and his colleagues have tackled a wide range of stories that could not practically have been done without AI — at least not without its power to search through thousands and thousands of documents, categorise them while benefiting from AI's ability to analyse language and meaning.
The stories have included a serious look at whether Europe was really rearming as claimed.
"A lot of money is being spent on armaments but is it really going into concrete and metal and new factories or is it being wasted," Cook asks.
The FT got publicly available radar satellite data from the European Space Agency and used it to survey 150 European defence industry sites to detect how much building had been going on over the past five years.
"It would be completely impossible without, not just AI, but using the infrastructure we have created to use AI," he says.
The story revealed a huge surge in rearmament across the continent with 7 million square metres of new factory and ammunition plant development.
The FT's AI team has even used the technology to see the impact AI is having on business in its own right.
They used the technology to extract over five years of references from 500 American companies tracking the boom years of AI and found a trend towards greater nervousness and caution. The FT also found that apart from the fear of "missing out", few of the big listed corporations could actually describe how the technology was changing their businesses for the better.
AI investigative journalism within the FT is still something of a minority preoccupation but Cook could hardly be more proud of the fact that the technology is starting to be used by the FT's foreign bureaux, in particular by Laura Dubois in Brussels.
Laura, a data savvy reporter used the FT's AI system to go through thousands and thousands of filings by MEPs on their second jobs.
The findings were dramatic, although not perhaps unexpected, that a lot of MEPs were not fully disclosing their second jobs and that a lot of those jobs demonstrated a conflict of interest with their duties as MEPs.
His journalistic journey
As for Cook, he has ended up as a ground—breaking journalist almost entirely by accident. There is no journalism in his family — his mother was a teacher and his father was in shipping. After a history degree and a masters in economic and social policy at Oxford, Cook was a hired as a Peter Martin Fellow at the FT as a leader writer — a fellowship named, appropriately, after the man who established FT.com who died tragically young in his mid—fifties.
"I hope Peter would be cheering what technology has become," Cook adds.
The FT journalist later went to Newsnight, essentially because he likes trying new things, when the then editor Ian Katz was looking for journalists from outside the BBC in the wake of the Jimmy Savile and Lord McAlpine scandals.
He notes that some television staff had problems accepting newspaper journalists.
"I was there for five years and you used to get these 23—year—old— producers who regarded you as tainted with less television experience. If you have ever worked on a newspaper you were always a newspaper person", Cook says.
He left in 2019 long before the cull of the programme's specialist reporters.
He is not a fan of what has happened at Newsnight.
"In newspaper terms, the BBC has really good places for 450—word stories and a lot of places where you can file a 3,000—word story but nowhere in between. Newsnight was taking in those size of stories and it was productive," Cook argues.
Newsnight also used to have its own reporters offering a degree of diversity. It is "quite dangerous" for all BBC journalists to be working directly for the director of news and current affairs because you can end up with one opinion.
Good and bad uses of AI in the newsroom
Cook is already very clear where the limits should be drawn on the use of AI.
"It is really good at searching and its really good at categorising but you have to be really careful to keep it in that box for now," explains Cook.
"It cannot exercise judgements and you have to be very, very diligent not to allow it to do that," says Cook who argues for the need for a very strong editorial guardrail system.
There is another matter of paramount importance, the FT journalist believes. You must never end up admiring the technology for its own sake and losing sight of the story.
"The single most important challenge for the newsroom in this stuff is to keep your eye on the story and forget about showing off how clever you are," he says.
Cook accepts that AI is very much a double—edged sword and gives bad actors the ability to generate thousands and thousands of words of "crap".
He thinks it would also obviously be a disaster if local newspaper publishers got rid of their reporters and replaced them "with rewriting machines".
In fact, Cook believes that smaller publications such as local and regional newspapers can use AI using easily available software on home computers.
"It is going to get easier to do cool new stuff even if you only have a few reporters," says Cook. He believes the technology can help editors in particular to stay on top of filings and court hearings while managing stories better.
Sometimes, Cook thinks that rather than being at the cutting edge of technology, there is an old—fashioned feel to what AI enables you to do.
It's like 19th century state of the nation reporting such as collecting data on the condition of the people as did Charles Booth, founder of the Salvation Army.
"I think that we can become more efficient using AI. That doesn't necessarily mean writing more but when we write, we can do things we couldn't do before," Cook explains.
AI can also greatly speed up more conventional reporting as Cook and his Mumbai—based colleague Krishn Kaushik demonstrated recently.
The two reporters revealed that steel tycoon Lakshmi Mittal's energy joint venture in India had stopped buying Russian crude oil after the FT demonstrated the company had bought almost $280 million of oil that was transported on sanctions—listed vessels. At the same time, the White House has been increasing tariffs to try to prevent India buying Russian oil.
The FT revealed the fact that sanctioned vessels took four shipments of oil as far as the Gulf of Oman before transferring the cargoes to tankers that were not covered by sanctions for the final leg of the journey to India.
Cook and Kaushik traced the customs filings that showed the oil had been bought from a St Petersburg oil supplier.
On the whole, Cook is positive about the increasing use of AI by journalists.
"The outlook for journalists to experiment in ways that are ethical is really bright and getting brighter," Chris Cook concludes.
And, of course, many such stories, which could not have been tackled by any other means, add considerable added value and help the Financial Times coverage stand out from the endless stream of online "breaking news".
This article was first published in InPublishing magazine. If you would like to be added to the free mailing list to receive the magazine, please register here.
