Regardless of whether you agree with him or not, there is something attractive about Dermot Desmond’s suggestion that we scrap MetroLink and put the money into artificial intelligence-based solutions to Dublin’s transport needs. The measure favoured by the billionaire is self-driving cars or autonomous vehicles, as they are known.
The idea that we can use technology to leapfrog a decades-old problem – such as Dublin’s transport woes – is in line with the techno-utopian narrative talked up by the likes of Mark Zuckerberg, Elon Musk and Jeff Bezos. This doesn’t make Desmond wrong, but for some people, it is all the reason they need to dismiss the idea out of hand.
At the same time, it would be a rash move for the Government to abandon a project (estimated to cost €23 billion) based on proven technology and instead bet the farm on an AI-based alternative that doesn’t exist. It’s a bit like El Salvador deciding to adopt bitcoin as legal tender in 2021. That did not work out too well despite all the crypto bros who said it was the way to financial stability.
The easiest argument – but probably the weakest one – to make against Desmond’s idea is the one outlining that the technology doesn’t actually exist. There may be no example of a city with a public transport system comprised of AI-controlled autonomous vehicles, but it’s not hard to see how there could be one developed and operating by 2035, which is when the MetroLink is due to start running.
READ MORE
The World Economic Forum – the people behind the annual Davos CEO gabfest – predicts that fleets of robo-taxis will be operating at scale in 40 to 80 cities by then. “China and the US are expected to dominate the roll-out in cities, with Europe and the Middle East also emerging opportunities,” they suggest.
There will also be a significant number of autonomous trucks in operation by then, although fully self-driving cars will take longer, they predict. Progress will be determined by consumer trust, prices, technology and regulatory obstacles as well as the need for scale.
You can get into a debate about Dublin’s capability to meet these criteria. Equally, one can question if the things holding back infrastructure projects – planning in particular – will also militate against the introduction of a public transport system in Dublin based on autonomous vehicles. But the salient point is that the technology will be a reality and Desmond’s suggestion is far from fanciful.

A more powerful argument against going down this route is a fundamental one about the power struggle between governments and technology companies that have their economies by the throat. It is currently playing out across all developed democracies as these behemoths seek to shape the regulatory and business environments in their favour, often with the willing assistance of politicians.
To date, none of these companies have demonstrated any conception of their responsibility to society and the need to exercise their power in the wider interest. Their overriding imperative remains generating bigger profits. That in turn reflects the libertarian ethos permeating the financial ecosystem that lies behind them.
There have been many books written about the nature of this hyper-capitalism and the innovation it has spawned. However, the bottom line is clear here. If you were worried about Transport for Ireland franchising out some bus routes in Dublin to private companies, you had better hang on to your hat if robo-taxis arrive.
A metro that should have been built 30 years ago is not the solution to Dublin’s transport problem, but it is part of it
The prospect of the entire Dublin transport system being in the hands of some sort of Tesla-cum-ChatGPT conglomerate owning and operating autonomous vehicles should fill you with dread.
You would have to seriously question the ability of the Government to oversee – and regulate in the public interest – an AI-based public transport system operated by entities sharing an ethical framework and sense of public accountability with big tech companies.
There isn’t much chance of them being hauled in front of the Oireachtas transport committee to perform a mea culpa over robo-ghostbuses.
Hopefully, the reality will be a little more prosaic. There is no real need for a binary decision between AI and MetroLink. A metro that should have been built 30 years ago is not the solution to Dublin’s transport problem, but it is part of it. AI will also be part of the solution.
The trains that we have been told will run every three minutes at peak times on MetroLink have probably not even been designed yet, but it seems inconceivable that AI will not be involved in managing and operating them. Likewise, the other transport services that will connect to MetroLink – and elsewhere – are all obvious candidates for the use of autonomous vehicle technology.
Desmond is neither right nor wrong about MetroLink versus AI, but what he has done is prompt us to raise our ambition for what can be achieved far above simply building a metro.