(CNN) Scientists using artificial intelligence to sift through around 100,000 climate studies were trying to put a number on how many people in the world were already experiencing the impacts of the climate crisis.
Instead, they discovered something else: there is a worrying inequality in the world of climate science.
Climate change studies are twice as likely to focus on wealthier countries in Europe and North America than low-income countries like those in Africa and the Pacific Islands. That blind spot is a problem, as the Global South is and will continue to be more profoundly impacted by the climate crisis than wealthier countries.
The ability to link the climate crisis to real-world impacts has grown dramatically in the past decade, as more people face the consequences of a warming planet, including deadly floods, destructive wildfires and crippling heat. But it has been a challenge to collect and scrutinize the vast amount of research, to fully understand the global impact.
In research published in the journal Nature Climate Change on Monday, scientists used machine learning — training computer algorithms to detect patterns and predict outcomes — to analyze more than 100,000 climate change studies.
"There's just so much climate science produced, like tens of thousands [of studies], and getting to grips with this evidence is really difficult," Max Callaghan, lead author of the study and a researcher at the Mercator Research Institute on Global Commons and Climate Change, told CNN. "So we trained the machine learning algorithm to predict the areas that we didn't have time to look at — which is most of them."
Compiling the results of all of those studies would suggest a vast majority of the world — 80% of land area, where 85% of the world's population lives — is experiencing the effects of the climate crisis right now. It's a large percentage, but experts know the true number is even higher.
The authors called the blind spot in research an "attribution gap." Callaghan said the gap suggests 85% is likely to be an underestimate.
Friederike Otto, co-lead of the World Weather Attribution initiative, who was not involved with the machine-learning research, also said the study's estimate is likely too low. Over the years, climate scientists like Otto have been saying the climate crisis will leave no place in the world untouched.
"The study focused on changes in mean temperature and precipitation, rather than extremes, but we know that heat extremes are changing faster than mean temperatures and that heat extremes are increasing almost everywhere," Otto told CNN. "It is likely that nearly everyone in the world now experiences changes in extreme weather as a result of human greenhouse gas emissions."
"What we find here is that the evidence is distributed unequally across countries," Callaghan said. "And this is really important because often when we try to make a map or to find out where the impacts of climate change are happening, we find often few scientific papers in less developed countries or low-income countries."
Callaghan added that this attribution gap leaves people wondering if climate change is happening in those areas, even though climate scientists firmly believe so.
"We want to try and point out that absence of evidence isn't evidence of absence," he said.
The authors note in the study that an automated approach is "no substitute for careful assessment by experts," however, it can identify large numbers of studies for a region that may point towards consequences brought by human-caused climate change.
Tom Knutson, senior scientist with the National Oceanic and Atmospheric Administration and co-author of the study, told CNN the machine learning methodology has "a number" of limitations and weaknesses, since it only accounts for certain climate impacts — in this case, human-induced precipitation and temperature changes.
If it accounted for other impacts such as sea level rise, for example, he said the outcome may have suggested "a greater fraction" of the world's population has been experiencing climate change.
A recent report by the World Meteorological Organization found that an extreme weather event or climate disaster has occurred every day, on average, somewhere in the world over the last 50 years, marking a five-fold increase over that period.
This summer alone was packed with extreme weather events across the Northern Hemisphere: While the United States has been battered by a cocktail of drought-fueled wildfires, devastating floods and a historic heat wave, China and Germany experienced deadly flooding events in July as southern Europe and Canada battled destructive wildfires of its own.
Despite the observed extremes, the dearth of substantial scientific evidence has the effect of limiting the changes that can be proposed or implemented in under-studied locations, Callaghan said.
"It's useful to bring together the literature and data as this study has, which allows us to see where more data is needed and where there are gaps," Otto said, pointing to previous studies. "Their finding of a gap in the Global South is similar to what we found last year, where we saw that extreme events are identified less often and are the subject of fewer attribution studies when they occur in poorer countries."
World leaders will gather at a critical UN climate meeting in less than a month, and one of the issues that will be discussed is the amount of funding developed nations can pledge to help the Global South move away from fossil fuels and manage the impacts of the climate crisis.
Callaghan said this new machine learning research delivers a key message for world leaders: Climate change is already happening and the planet will only continue to warm, meaning adaptation is critical as well as halting the use of fossil fuels. The new study provides an outline of where more climate funds and climate research are needed, and it is up to global leaders to implement that.
"The world will continue to warm until we stop burning fossil fuels, and there is simply no way around that," he said. "And what we really need to recognize is that we need to change the trajectory and reduce emissions."
Correction: This story has been updated to reflect the study was published in the journal Nature Climate Change.