Gallese: Banks are removing humans from risk assessment claiming computers can do the job. They are wrong.
That distinction matters.
According to Morgan Stanley, up to 10% of the workforce across major European banks could disappear by 2030.
The cuts won’t hit traders or CEOs.
They’ll hit:
– back-office
– risk management
– compliance
– internal controls
The parts of the bank most people never see.
And that’s exactly the problem.
It's different from “AI replacing jobs”. This is management replacing institutional judgement.
Let me explain why that’s a systemic risk.
1. Risk functions exist to slow decisions down
Risk management and compliance are not about speed.
They exist to:
– question assumptions
– amplify doubts
– force humans to explain decisions
Automation flips that logic.
AI systems optimise for:
– speed
– stats
– pattern recognition
Not moral judgement.
Not context.
Not dissent.
And not compliance.
When all of this is automated, judgement becomes optional.
2. Automation erases the learning pipeline
A JPMorgan executive warned about this explicitly:
If junior bankers never learn the fundamentals, the system loses its memory.
Literally.
Risk expertise is apprenticeship-based:
– you learn by reviewing edge cases
– by seeing mistakes
– by staying updated
– by watching senior judgement under pressure
If AI absorbs that layer, future leaders never learn how risk actually forms.
You don’t just lose jobs. You lose institutional understanding.
Dr Chiara Gallese is on LinkedIn
3. Efficiency gains hide delayed failures
AI can seem to improve operational efficiency by 30%.
That’s their selling point.
But risk management failures don’t appear immediately. They compound slowly.
Automation makes systems *appear* safer:
– fewer alerts
– no pushbacks
– faster approvals
Until one assumption breaks.
And when it does, fewer humans will know how to intervene.
4. This is the same pattern we keep seeing in AI deployment
Different sector. Same playbook:
– risk management automated
– human oversight reduced
– escalation paths flattened
– accountability pushed downstream
– safeguards follow deployment
We saw it in:
– content moderation
– algorithmic advertising
– automated decision-making
Banking is just the next critical infrastructure layer.
AI is not the problem here, it's just an excuse for bad decisions.
And we must look further:
Who will suffer the consequences when AI-driven risk management fails?
When:
– models approve what humans would flag
– compliance becomes a checkbox
– new risks ate ignored
– and no one remembers why controls existed in the first place
European banks are re-architecting how risk is understood, challenged, and mitigated.
Once that shift is normalised, rebuilding human judgment is far harder than automating it.
And clients will bear the costs.
- Editor's note: Dr. Gallese uses the common term "AI" which WMLR and its associated activities reject because machines cannot be intelligent. It's been included here for the sake of continuity.
About this section
Opinion pieces or "Op-Eds" are the home-made bombs of the publishing world. So long as they meet editorial standards, are not intentionally offensive with a view to causing hurt or insult and are relevant to our field of endeavour, we will look at submissions.
We like contentious, we like contrarian views. We don't like pretty much any -ism . We recognise that Opinion pieces are one person's view and are not balanced (if they are balanced and reach a reasoned conclusion, they are probably more suited to the Articles section). We do not like acronyms and buzzwords.
Op-Eds are the author's personal views and do not necessarily represent the views of World Money Laundering Report or its publishers.
To submit an Opinion piece, please complete the Contact form.
