Host: Artificial intelliArtificial intelligence is no longer just a tool being tested in one corner of a company. More and more it is connected across entire organisations and it is making real decisions — especially in financial services. That raises a big question: who is responsible when something goes wrong? Europe has written some of the world’s toughest rules on AI. The US and the UAE are taking very different paths. But do more rules actually make AI safer? Joining us now is Michael O’Loughlin, USA Ambassador for the MENA Fintech Association and Managing Partner at Argonautic Global. Michael, thanks for being with us again today.
Really good to be back.
Europe has led the way on AI rules. Why does it not seem to be leading the global conversation on AI safety?
As an Irishman, I would be pushing back a little on that premise. There are three different models taking place. Europe has really leaned into the conversation and rule-making to set the guardrails — those conversations and debates are very important. The UAE has taken a different approach that is quite admirable — they have aligned the government, capital, and industry, which is often the most difficult thing, and they did it at pace. And in the US there is huge commercial pressure — go to market, win, win, win. They are all going to have three different results. But if they could take a step back, they could learn from each other.
Are we confusing regulating AI with making it safe?
You should not run with scissors — but you certainly should not do it with the lights off. With regulation you can be compliant, and that is a very good thing. But regulation does not wake up at 2 a.m. if something goes bump in the night. In banks, AI is being supplied by different vendors and used by different verticals — compliance, treasury, risk, fraud — and they are all taking information from each other. The coordination piece is super important. Who owns it if something goes wrong? That comes back to accountability — something I keep coming back to.
What are you seeing companies do differently across Ireland, the UAE, and the US when they put AI to work?
In Europe — if you fail to prepare, you prepare to fail. They really want to focus on the foundation. It is not about speed. They are looking at it holistically — how do we attract foreign direct investment and build something that will last? In the US, a lot of people are going to be seeking forgiveness rather than permission. But if there are no rules at the moment, that permission piece is up for grabs.
When AI makes a big financial decision that goes wrong — who is accountable?
You cannot send a bot to prison. Accountability needs to be human. It needs to be an organisation — one throat to choke. Someone who will hold their hand up and say if this goes wrong, whether it is a vendor. Contracts are going to look very different with vendors going forward. Regulators are going to have very different relationships with banks. The EU AI Act coming out very soon will set a new standard — and I think the coordination from the UAE and Europe will certainly impact what is happening in the United States.
What do C-suite executives actually say when you ask who is responsible?
Success has many fathers and failure is an orphan. Regulation is not good until things go wrong. I was talking with some C-level bankers near the stock exchange just yesterday. It is not keeping them up at night yet — but yet is the important word. Their question was: when do you think regulation will come in here? Because there will be fines with regulation. And this is something they need to be prudent about.
For the average person watching — what should they know about how AI is already making decisions about them?
Going back to my Citibank days 20 years ago, we were always building customer-centric models based on data people provided. Now banks are looking beyond what you have given them — monitoring social media, monitoring location. Anticipating through algorithms what you will need next. On the retail side, if you have a family of three or four, they know at some point you will need to save for college or a car loan. On the corporate side, AI is learning what the data flows around the world are, what the payment corridors are. Looking at West Coast startups — the AI hallucination gap is getting smaller and smaller. We are still in infancy with AI and finance, but I think we are going to make a giant leap in the next 18 months.
In five years — who wins: the company that moves fastest on AI or the one that manages it best?
My hope is the ones that manage it best. Safety first. It is not going to be one model that wins. There will be multiple models and the institutions will choose — and hopefully the end users will be serviced properly by the larger institutions as a result of those choices.
Michael, this is fascinating. Thank you so much for joining us.
It is super interesting. Thank you.
