When summarizing scientific studies, large language models (LLMs) like ChatGPT and DeepSeek produce inaccurate conclusions in up to 73% of cases, according to a study by Uwe Peters (Utrecht University) and Benjamin Chin-Yee (Western University, Canada/University of Cambridge, UK). The researchers tested the most prominent LLMs and analyzed thousands of chatbot-generated science summaries, revealing that most models consistently produced broader conclusions than those in the summarized texts.
Prominent chatbots routinely exaggerate science findings, study shows
More In Finance
-
Canada has some of the highest interchange fees in the world. Interchange fees are the fees businesses pay each time their customers pay by credit card. The average interchange fee in Canada is about 1.5 per cent of [...]
-
Main Street businesses that survived COVID-19 restrictions are now navigating a pandemic recovery where predicted changes in the retail industry have been accelerated by five to 10 years. The ability to adapt to these changes, coupled [...]
-
The big idea Consumers who see a product on sale being virtually touched are more engaged and willing to pay more than if the item is displayed on its own, according to a recent research paper [...]
-
Entrepreneurs, their associated startups and the subsequent growth of their companies have a vital impact on the health of our economy. In Canada, young adults have demonstrated a growing interest in entrepreneurship. Entrepreneurship has historically been narrowly [...]
-
Economics is broadly divided into macroeconomics and microeconomics. The big picture, macroeconomics, concentrates on the behavior of a national or a regional economy as a whole: the totals of goods and services, unemployment and prices. Then there’s a more [...]
