As someone working in a profession very likely to be substantially affected by artificial intelligence, the latest news from OpenAI and Meta that their new large language models will soon be able to reason sent a chill down my spine.
On the one hand, what fascinating progress. The positive implications for the quick flagging of potential financial crime, for example, should be welcomed. Not to mention wider applications outside of finance, like in healthcare diagnostics.
On the other hand, the livelihoods of journalists, mine included, may well be wiped out in a not-at-all-distant future.
“We are hard at work in figuring out how to get these models not just to talk, but actually to reason, to plan … to have memory,” Joelle Pineau, vice-president of AI research at Meta, told the Financial Times. Meta is launching its Llama 3 model in the coming weeks, and GPT-5 would be available “soon”, according to its creator, OpenAI, which is backed by Microsoft.
Are bankers beginning to worry about their own jobs, too?
I asked one of my most colourful sources. “I haven’t heard anyone at [my bank] expressing anxiety about losing their jobs due to AI… [This is] probably all ignorant and naive,” they said.
An entrepreneur attempting to shake up banking (and working with a growing number of big, international banks) has a more decisive view: “AI is to banking what robotics was to the auto industry… Banks will die, but banking will blossom.”
While both of my contacts would have likely toned down their views in public, their comments point to a potential existential threat for bankers. How will their jobs look next year?
As I leave you with that question (and hope more of you will share your concerns and insights with me), others are looking at the even more imminent consequences of AI in the analysis of banks’ own performance, as well as that of any company releasing financial accounts.
You may recall Claire Bodanis, the UK consultant on a mission to regulate the use of AI in corporate reporting. I wrote about her here. Specialists in the EU are also paying attention to the matter — the Brussels-based European Financial Reporting Advisory Group recently invited Bodanis to discuss her concerns with its staff.
The latest AI news has added urgency to Bodanis’s argument. Models that are able to reason may present commentary that is “potentially becoming more plausible, but not necessarily right”, she says.
With narrative reporting (the part of the annual statement that describes the company’s business model, performance and strategy) being “largely opinion”, as Bodanis puts it, the potential use of AI to populate such information can be problematic. “Narrative reporting is … the management’s analysis of the data,” she says, and it could be tricky if the conclusions of such an exercise were based on reasoning AI.
Further, if using such new technologies may be tempting to time-poor managers, and the work could be competently done by a machine, Bodanis asks: “What’s the management team for? This is potentially a threat to the more strategic jobs”.
I’m keen to hear your thoughts on this — both if you feel your job is at risk, or if you’re working on developments that may make those strategic jobs redundant…even, and particularly, if it is something directly affecting journalists’ jobs.
Silvia Pavoni is editor in chief of The Banker. Email her at silvia.pavoni@ft.com. Follow her on LinkedIn here.
