Asset managers pore over central bank communiqués as if they were literary critics analysing poetry. The officials, meanwhile, are determined to manage expectations to prevent panics as they taper and tighten.
Central bankers hold conferences and write papers on communication strategies. The finest economists in the world look at a topic as far away from their specialism as the Flesch-Kincaid score of language difficulty. They also stress that they want to listen to the market rather than just speak in an echo chamber.
A couple of years ago the Financial Times compared this intense, artificial and ultra-sensitive relationship between investors and the US Federal Reserve to the psychological thriller Gone Girl, where a couple is riven by mutual mistrust after the sociopathic wife fakes her own murder.
But now central banks are preparing to deal with a new partner, whose motives and rationale could be harder to understand.
Benoît Coeuré of the ECB mentioned in a speech at the Bank of France on Friday that in the future central bankers may adapt what they say for an audience of robots.
“We may one day be tempted to draft our monetary policy statements and speeches in the light of how they will be comprehended and interpreted by artificial intelligence algorithms," he said. "So in the future big data may work two ways, with central banks acting on, and reacting to, it, with consequences which remain to be understood.”
If central banks are to start having to 'communicate’ with algorithms, it could be more reminiscent of the film Ex Machina, where in an equally destructive relationship, an engineer is seduced and betrayed by a robot trying to escape from captivity. Both have to work out whether to trust each other. Ultimately it is the robot who is bluffing and leaves him trapped.
One of the benefits of artificial intelligence is that it would bring into being rational actors not prone to hidden biases or herd mentality. But for all the tech utopianism, no one knows how AI risks will evolve.
Interdependency and interconnectedness of AI services could amplify market shocks, a study by the Financial Stability Board (FSB) warned earlier this month. Three human investors can listen to an ECB statement and react in different ways, due to their biases. Three robot investors powered by the same technology may tend towards the same interpretation. And that interpretation could be harder to judge, given what the FSB termed AI's "lack of interpretability or auditability".
One answer could be to go the whole hog and automate communication in central banking, making both participants in the relationship robotic. Who better to speak to a robot than another robot?
Some wags might suggest sitting through an ECB press conference is rather like listening to robot already.
But there is still a human touch necessary, because not all stakeholders in central bank decisions are machines. As Christine Lagarde said in September at a Bank of England conference: “Good monetary policy, as we know, is about story telling... Can machines really be held accountable to the young couple unable to buy a house, to the working mother finding herself unemployed?”
Politicians have loved attacking central banks recently. US president Donald Trump featured Fed chairperson Janet Yellen in a campaign advert ahead of the 2016 election to pit himself against “global special interests”. Brexiteers have bashed Mark Carney for supposedly doom-mongering about the UK leaving the EU.
In the face of scrutiny and threats to their independence, central bankers need a human face now more than ever. Editing comments for the benefit of computer programme should be way down the list of central bank communication priorities.