Artificial intelligence is expected to play a decisive role in the future of fundamental physics research and in unraveling the deepest mysteries of the universe, according to CERN’s incoming director general.
British physicist Professor Mark Thomson, who took over the leadership of CERN on January 1, 2026, says machine learning is opening up entirely new possibilities in particle physics, comparing the development to the AI-driven breakthrough in protein structure prediction for which scientists at Google DeepMind won the Nobel Prize in Chemistry last October.
At the Large Hadron Collider (LHC), these same techniques are already being used to detect extremely rare phenomena that may be important for understanding how particles acquired mass after the Big Bang, and for exploring theoretical scenarios for the long-term stability of the universe itself.
New accelerator on the horizon
Thomson’s remarks came as CERN’s board of directors moves forward with plans for the future Circular Collider, a proposed new accelerator with a circumference of 90 kilometers that would significantly exceed the LHC. Despite some skepticism after what critics describe as the lack of dramatic discoveries since the Higgs boson was identified in 2012, Thomson believes AI is breathing new life into exploring physics beyond the Standard Model.
He predicts that major advances could occur after 2030, when LHC upgrade plans will increase the intensity of the particle beam by a factor of 10. This will enable unprecedented observations of the Higgs boson (commonly known as the “God particle”), the particle that gives mass to all other particles and defines the field that holds the structure of the universe together.
God particles meet themselves
“There are certain measurements of the Higgs boson that are fundamental to the nature of the universe,” Thomson explained. “We will try to produce not one but two Higgs particles at the same time.”
This will allow scientists to measure for the first time how the Higgs boson gives itself mass, a phenomenon known as Higgs self-coupling. Although it is extremely rare for two Higgs particles to be produced at the same time, Thomson said he was now confident that it was achievable. “Five years ago, we would have thought this was out of the reach of the LHC. Now we are confident that we can get reliable measurements.”
The strength of this self-bond is important for understanding how particles first gained mass a trillionth of a second after the Big Bang. It could also reveal whether the Higgs field remains stable or could undergo new transitions in the future, a scenario that could result in the extinction of the universe as we know it. The Standard Model shows that this is theoretically possible, but there is no immediate cause for concern.
CERN theoretical physicist Dr Matthew McCullough said: “That’s not something that could happen on a time scale even relevant to our star.” “But it remains a purely scientific question: Could it happen?”
Thomson said this is a “deep and fundamental property of the universe that we don’t yet fully understand,” adding that if the measurement of self-association deviates from theory, it would be a “monumental discovery.”
AI at every step
Artificial intelligence is now being applied at every stage of LHC operations, from the selection of data to be collected to the subsequent analysis. “When the LHC collides protons, there are about 40 million collisions per second. We have to decide within microseconds which events are worth keeping,” explained Dr. Catherine Rennie of the ATLAS experiment. “Thanks to AI, we are at least 20 years ahead of what we originally expected to achieve.”
Scientists also hope that the LHC may one day produce dark matter, which is thought to make up most of the universe. Its nature remains unknown, making exploration inherently difficult. Generative AI could help by enabling more open-ended research approaches. Mr Thomson said: “Instead of searching for specific signals, you can ask if there is anything unexpected in this data.”
