AI students get heads together in Brussels to fight flood risk

Students in AI have gathered in Brussels for a competition to develop models that can be used to mitigate future flooding disasters.

Their work aimed at potentially saving lives follows global alarm bells over AI's potential threat to humanity.

So how does this generation of AI developers view this divisive technology?

Belgian data student Simeon Godfroi is immersed in artificial intelligence. 

"It's frightening," he says. "Even more frightening for me given I know the extent to which AI can impact society."

That fear has been a tipping point for AI after its ability to escape human control triggered alarm bells over its potential threat to humanity – bells rung particularly loudly by Jacob Coxon, who in September resigned from Anthropic while earning that AI "could kill us all by the end of the decade".

Coxon isn't much older than Godfroi, who's among 60 students gathered in Brussels to pool their AI expertise for building flood warning and prevention technology.

A two-day competition organised by the FARI research center – whose motto is 'AI for the common good' – and the European Commission is looking for the best models that can be applied to communities at risk of flooding.

They're using data from deadly floods caused by record rainfall that swept through parts of Belgium, Germany and the Netherlands in 2021.

Just before this gathering, parts of Europe were battered by torrential rain and flooding – and scientists warn climate change is increasing their frequency and potency.

At the same time, world leaders have been shaken by multiple cases of evidence of AI going rogue, raising concerns it can be weaponized.

One of the competitors, law student Alessandra Soto Ruiz, thinks regulation is overdue but the race to develop the most advanced AI models among competing global powers means applying the brakes is difficult.

"It's a field that needs to be regulated," Soto Ruiz says. "But there needs to be a space for Europe to grow its foundational models, its technology, so it is a difficult field to regulate also because it's an evolving sector."

Even so, not all of these AI exponents are convinced by the potential risks.

"It is not yet an existential threat," says AI doctoral student Gaspard Merten.

"Yes, it has been improving, it improves a lot. But will it continue to improve? No one can tell, it's impossible to say. We can forecast the future a bit, but we cannot say it will become sentient or anything yet."

Despite these concerns, flood prevention is a critical area for AI, with leading exponents like China investing heavily in advanced mitigation tools that can identify the most vulnerable areas and assess gaps in emergency responses.

For Carl Mörch, FARI's Co-Director organisers FARI, bringing the AI generation together to work in teams is a reminder that for all its progress, this rapidly evolving technology still needs human stewardship.

"With all this information, with all these possibilities, you need to make the right choices, you need the right data, the right evidence. So you also need proper choices made by humans."