Wisdom of AI Crowds

What if you held a remote control for the world? Imagine being able to pause a major national crisis, rewind to change a critical decision, or press fast-forward to see the exact long-term consequences of your actions fifty years into the future.

In the real world, policymakers face the exact opposite reality. When trying to
address the biggest societal problems such as the global climate change
phenomenon or localised land-use disputes, there is no dress rehearsal.  These problems are called “wicked problems” in public policy. Once a policy is introduced to address these issues, they are regarded as “one-shot”.

Remember the recent issue of Manoel Island? Following intense public pressure and NGO campaigns, the government moved to turn the area into a public park.

Yet, just a few months later, an unexpected controversy erupted over the sudden appearance of commercial padel courts on a disused sports ground.

Who could have predicted that a policy meant to save a green space would almost immediately spin off into a new battle over commercial layout?

This is what textbooks call a "wicked problem", a challenge so complex, dynamic, and deeply polarised that solving one part of the puzzle almost always creates a new, unpredicted problem somewhere else. Because humans cannot foresee the future, it is nearly impossible to map out how different interest groups will react to a policy change in the real world.

To break this impasse, groundbreaking doctoral research by Gian Paul Gauci at the Department of Policy, Politics and Governance of the University of Malta is
constructing that virtual remote control through a new framework called the Wisdom of AI Crowds.

By building advanced "Artificial Societies," we create a safe, digital twin of complex policy environments. We populate this virtual sandbox with autonomous, data-grounded AI personas that accurately represent the real-world norms, values, and beliefs of actual stakeholders, from developers and activists to state authorities and ordinary citizens.

When a new law or regulation is introduced into this virtual world, we can fast-
forward time to watch how these conflicting groups interact, compete, or build
alliances over months and years. By running these scenarios thousands of times and checking the results with human experts, we can safely discover which policies build true public consensus before a single real-world action is taken.

The goal is not to hand our democratic decisions over to machines, but to give our leaders a risk-free laboratory to stress-test ideas, clear political gridlocks, and protect our environment for generations to come.